Copyright TechTarget - All rights reserved https://cyber.law.harvard.edu/rss/rss.html Techtarget Feed Generator en Sun, 19 Apr 2026 17:23:44 GMT https://www.techtarget.com/searchaws editor@techtarget.com <p>AI agents are autonomous intelligent software components that form the foundation of artificial intelligence (<a href="https://www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence">AI</a>) systems. Agents are designed to perform specific tasks independently without the need for human intervention.</p> <p>Intelligent agents are conversational and can interact with other systems, such as applications and APIs, access data, perceive specific environments, exercise reasoning, make decisions, take actions to achieve defined goals and learn from prior outcomes to refine future decision-making. These capabilities help organizations become more productive by delegating repetitive and mundane tasks to these AI agents and freeing human resources for more complex and strategic activities.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/5eQwl9yaEjE?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <section class="section main-article-chapter" data-menu-title="How do AI agents work?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How do AI agents work?</h2> <p>AI agents use machine learning (<a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-ML">ML</a>) and techniques such as natural language processing (<a href="https://www.techtarget.com/searchenterpriseai/definition/natural-language-processing-NLP">NLP</a>) to take on a range of tasks, from simple queries to complex problem-solving. Unlike traditional AI, AI agents can self-learn and continuously improve their performance.</p> <p>These agents follow a cycle of perception, reasoning and action or outcome. This is sometimes expressed as sensing, thinking and acting. An agent's workflow typically defines the goal based on user input, breaks it into smaller subtasks capable of accomplishing the intended goal, and executes those subtasks using production data, such as inputs from IoT devices; knowledge base, such as external data sources; the Web; and other tools.</p> <p>The following are the operational steps AI agents take:</p> <ol class="default-list"> <li><b>Define goals.</b> The process is initiated when an autonomous AI agent receives precise instructions or goals from a user prompt. These instructions act as the cornerstone for the agent's subsequent actions.</li> <li><b>Gather and process data.</b> An AI agent gathers instructions and discovers and processes data through various <a href="https://www.techtarget.com/iotagenda/definition/smart-sensor">sensors</a>, inputs or data sources. This is the perception or sensing phase. For example, an autonomous car uses sensors to collect data about the road, traffic and obstacles, while an AI <a href="https://www.techtarget.com/searchcustomerexperience/definition/chatbot">chatbot</a> collects user queries.</li> <li><b>Organize and plan tasks.</b> The AI agent breaks the goal into smaller, actionable tasks to ensure efficient and effective task execution. This is the reasoning phase of operation. A single AI agent can handle these subtasks, or they can be delegated to other subagents that can provide specialized results.</li> <li><b>Execute tasks. </b>In this step, the agent takes uses various tools and techniques, such as using a large language model (<a href="https://www.techtarget.com/whatis/definition/large-language-model-LLM">LLM</a>), to automate tasks and manage complex cognitive activities. Execution delivers an outcome, such as using an actuator to implement changes in the real world, or making a decision or recommendation.</li> <li><b>Seek external feedback.</b> Once a task is executed, it's removed from the list, and the agent moves on to the next task. To assess progress toward the ultimate goal, the agent seeks external feedback and reviews its logs. This is the learning phase. During this process, additional tasks might be generated and executed to achieve the desired outcome. The agent also refines its decision-making to enhance future outcomes.</li> </ol> </section> <section class="section main-article-chapter" data-menu-title="What is the agent function and program in an AI agent architecture?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What is the agent function and program in an AI agent architecture?</h2> <p>AI agent architecture is a structured framework that enables <a href="https://www.techtarget.com/searchenterpriseai/definition/agent-intelligent-agent">intelligent agents</a> or systems to perceive, reason and act autonomously in their environments. The architecture can either be a physical architecture of a robot, such as actuators, sensors, motors and robotic arms, or a digital one, such as software agents or <a href="https://www.techtarget.com/whatis/feature/AI-content-generators-to-explore">content generators that use text prompts</a>, application programming interfaces and databases to enable autonomous operations. Both agent function and agent program are the main components that form the backbone of AI agent architectures.</p> <h3>Agent function</h3> <p>The agent function defines how an AI agent responds to its environment. The term <i>function</i> represents the ideal or theoretical response to every possible situation. For example, the agent function maps the agent's perceptions or the <a href="https://www.techtarget.com/searchenterpriseai/post/AI-agents-are-only-as-smart-as-the-data-that-feeds-them">data it receives from its environment</a> to actions. Before designing the agent function, most developers evaluate the required information, AI capabilities, knowledge base, feedback mechanisms and other necessary technologies.</p> <h3>Agent program</h3> <p>An agent program builds, trains and puts the agent to work on a chosen system, bringing the agent to life. The term <i>program</i> is the actual AI code implemented on a hardware infrastructure. It ensures the agent performs as intended, meets technical standards and operates efficiently. Although the program should precisely embody the function, there might be practical limitations or exceptions.</p> </section> <section class="section main-article-chapter" data-menu-title="How can AI agents be used?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How can AI agents be used?</h2> <p>Various industries use AI agents to enhance processes and automate tasks. These agents harness generative AI to assist with and collaborate on tasks, empowering users in the process.</p> <p>The following are examples of AI agent use cases:</p> <ul class="default-list"> <li><b>Customer support.</b> AI agents have diverse capabilities, such as responding to inquiries, managing refunds and providing advanced technical support. As a result, they're increasingly replacing traditional customer service chatbots. AI agents let businesses offer around-the-clock assistance without human intervention, improving customer satisfaction and reducing operational costs. AI agents are also integrated into apps and websites to serve as virtual AI assistants that enhance the customer experience.</li> <li><b>Finance.</b> AI agents are automating routine tasks, such as risk assessment and transaction processing, transforming the finance industry. By analyzing vast data sets, these agents provide insights that help drive the strategic decision-making of financial operations.</li> <li><b>Healthcare.</b> AI agents can handle and streamline healthcare-related tasks, such as answering inquiries, scheduling appointments, reviewing insurance, generating medical summaries and approving care requests. They also can analyze biological data and predict the efficacy of new drugs, accelerating drug discovery. Additionally, AI agents are able to personalize treatment plans, manage records and match patients to clinical trials. These capabilities help providers deliver better care and improve outcomes. Multiagent systems are especially effective for solving healthcare problems.</li> <li><b>Manufacturing.</b> AI agents can streamline and automate manufacturing tasks, such as quality control, <a href="https://www.techtarget.com/searcherp/feature/Predictive-maintenance-Definition-benefits-example-strategy">predictive maintenance</a> and <a href="https://www.techtarget.com/searcherp/tip/Use-cases-for-machine-learning-in-the-supply-chain">supply chain optimization</a>, . For example, AI agents can analyze real-time data to identify potential issues, optimize production schedules and improve product quality.</li> <li><b>Insurance.</b> In the insurance industry, AI agents are used to automate claims processing, risks assessments and responses to customer inquiries. By analyzing large volumes of data, these agents provide personalized policy recommendations, detect fraud and streamline administrative processes.</li> <li><b>Autonomous vehicles.</b> AI agents enable autonomous vehicles to operate with limited human intervention. These intelligent systems perceive the vehicle's surroundings and make informed decisions, such as when to turn or brake. By using AI sensors, AI agents detect stop signs, navigate unfamiliar terrain and adapt to changing environmental conditions.</li> <li><b>Smart environments.</b> AI agents automate buildings, as well as broader urban environments. Intelligent sensors in buildings can detect and identify people to unlock doors and ensure that the workspace is lit and comfortable. When the workspace is empty -- especially after hours -- the agents dim lighting and limit the use of heating or cooling to conserve energy. At the municipal level, agents collect IoT data about road and traffic conditions, optimizing traffic controls and street lighting to streamline traffic and save energy.</li> <li><b>Workplace automation.</b> AI agents can automate routine business processes, letting employees focus on higher-value tasks. For example, these agents can ingest documents, automate data entry, handle scheduling, and other repetitive and administrative tasks to smooth operations and boost productivity.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="Benefits and limitations of AI agents"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Benefits and limitations of AI agents</h2> <p>AI agents have numerous benefits and certain limitations. Balancing their advantages and drawbacks is essential for organizations seeking to utilize them effectively.</p> <h3>Advantages of AI agents</h3> <p>AI agents provide the following benefits:</p> <ul class="default-list"> <li><b>Increased efficiency.</b> AI agents automate repetitive tasks, such as answering customer inquiries, scheduling appointments and processing claims. This provides workflow automation and frees human workers to focus on more complex tasks. These capabilities also can reduce operational costs by automating many time-consuming tasks that are often subject to human error.</li> <li><b>Enhanced decision-making.</b> AI-powered agents use <a href="https://www.techtarget.com/whatis/definition/machine-learning-algorithm">ML algorithms</a> to analyze vast amounts of data quickly, providing deeper and more valuable insights that help businesses make informed decisions. This leads to innovation, such as automated research and development in data-heavy pursuits, such as biopharmaceutical research and software development and testing.</li> <li><b>Improved accuracy.</b> AI agents follow predefined rules, observe guardrails and learn from large data sets that minimize mistakes caused by fatigue or bias. As a result, AI agents reduce human error and improve task accuracy. Additionally, by analyzing patterns and making data-driven decisions, agents enhance the accuracy of certain tasks, such as data entry, diagnostics and financial analysis.</li> <li><b>Personalization. </b>AI agents analyze individual preferences and behaviors, delivering personalized experiences. For example, they can provide tailored buying recommendations in retail settings and customized treatment plans in healthcare.</li> <li><b>High-quality responses.</b> AI agents collaborate with other agents, use external tools and learn from their interactions. As a result, they provide more comprehensive, accurate and personalized responses compared to traditional AI models, leading to better customer experiences. It's important to note that these behaviors emerge naturally and aren't preprogrammed.</li> <li><b>Learning and adaptability. </b>Many AI agents learn and adapt over time to improve their performance based on feedback and new data, which leads to better outcomes.</li> </ul> <h3>Limitations of AI agents</h3> <p>The downsides of using AI agents include the following:</p> <ul class="default-list"> <li><b>Limited understanding.</b> Most AI agents rely on predefined rules, limiting their ability to handle complex or nuanced situations that require a deeper understanding of context. Although many AI agents can learn, complex exceptions require human direction to guide a suitable outcome or decision.</li> <li><b>Context and hallucinations. </b>AI agents can render outcomes that are inadequate or outright wrong. Limited context windows can fill, causing older context elements to be lost and allowing the AI agent to lose the underlying point of the user's prompt or inquiry. The AI agent can also hallucinate, creating false information that's presented as factual. Both issues must be addressed through ongoing AI agent design and development.</li> <li><b>Issues with adaptability.</b> While some AI agents can learn and adapt, their adaptability is often limited to specific environments or tasks, as they might struggle in dynamic or unpredictable situations. Vertical AI agents are emerging that are designed and trained to handle detailed and nuanced situations in specific industries, such as healthcare or law. AI agents are also used in combination to build orchestrated agentic AI workflows.</li> <li><b>Ethical issues.</b> Deep learning models can produce <a href="https://www.techtarget.com/searchenterpriseai/feature/6-ways-to-reduce-different-types-of-bias-in-machine-learning">biased or inaccurate results</a> due to insufficient or inaccurate data and bias in the underlying algorithms. Human oversight and clear explainability are essential to safeguard the output of AI agents, mitigate these risks and ensure fair and helpful responses. This has a direct effect on an organisation's reputation and compliance.</li> <li><b>Initial investment costs.</b> Although AI agents can lead to long-term savings, the initial investment in technology and training of the AI agents can be significant. It can deter some organizations from building and adopting them.</li> <li><b>Infinite feedback loops.</b> AI agents get into infinite feedback loops, where an agent's actions unintentionally trigger a chain reaction that loops back to the original action, creating an endless cycle. For instance, an AI agent designed to optimize a system might execute a change that worsens performance instead of improving it, leading to a series of adjustments that exacerbate the problem.</li> <li><b>Security and privacy concerns.</b> The data AI agents use often involves sensitive personal or business information. This raises concerns regarding <a href="https://www.techtarget.com/searchsecurity/definition/data-breach">data breaches</a>, misuse and privacy violations.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="The different types of AI agents"> <h2 class="section-title"><i class="icon" data-icon="1"></i>The different types of AI agents</h2> <p>AI agents can be classified into various types based on their characteristics, functionalities and the complexity of tasks they handle. Common types of AI agents include the following:</p> <ul class="default-list"> <li><b>Simple reflex agents.</b> These are the simplest agent types that operate on a set of predefined rules and don't possess any memory or the ability to learn from past experiences. They respond to stimuli in their environment and make decisions based solely on the current situation. For this reason, they're most suitable for straightforward and simplistic tasks.</li> <li><b>Model-based reflex agents.</b> Unlike simple reflex agents, model-based reflex agents maintain an internal state that reflects the environment's current situation. This lets them make informed decisions by considering both current and past inputs and adapting to changes.</li> <li><b>Goal-based agents.</b> Also known as <i>rule-based agents</i>, these AI systems have enhanced reasoning capabilities. In addition to evaluating environmental data, they compare different approaches to achieve the desired outcome. Goal-based agents always select the most efficient path and are well-suited for complex tasks such as NLP and <a href="https://www.techtarget.com/whatis/definition/robotics">robotics</a> applications.</li> <li><b>Utility-based agents.</b> These agents use utility functions to make decisions. They pursue goals and prioritize outcomes based on their perceived value. By evaluating the desirability of different states, they choose actions that maximize overall utility, making them suitable for complex environments where tradeoffs are inevitable.</li> <li><b>Learning agents.</b> These agents learn from experience and past interactions, improving their performance over time. They use ML techniques to adapt to new situations, refine their decision-making processes and become more effective in completing tasks. For example, a <a href="https://www.techtarget.com/searchcustomerexperience/definition/virtual-assistant-AI-assistant">virtual assistant</a> can learn about a customer's preferences and enhance its customer service capabilities.</li> <li><b>Hierarchical agents.</b> These agents follow a hierarchical structure, where higher-level AI agents program and direct lower-level agents to work toward a shared goal. This setup lets businesses break complex, multistep processes into simpler tasks, with each AI agent focusing on a specific set of responsibilities.</li> <li><b>Multiagent orchestration.</b> Although AI agents can operate individually, multiple agents can be combined to share information and collaborate to achieve more complex business goals. This interaction of different AI agents is typically orchestrated to create detailed agentic AI workflows.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="How to effectively implement AI agents"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How to effectively implement AI agents</h2> <p>AI agents have evolved beyond virtual assistants, such as <a href="https://www.techtarget.com/searchmobilecomputing/definition/Siri">Siri</a> and Alexa. They're proving valuable in fields such as drug discovery, fraud detection and supply chain optimization.</p> <p><a href="https://www.techtarget.com/searchenterpriseai/tip/How-to-prepare-your-business-for-agentic-AI-adoption">Setting up AI agents effectively</a> requires a strategic approach that covers planning, design, coding, implementation, deployment and monitoring phases. The following key steps help ensure successful execution of AI agents:</p> <ol class="default-list"> <li><b>Define clear objectives.</b> Before setting up the AI agents, companies should identify the goals they want them to achieve. Whether it's automating workflows, improving customer service or enhancing decision-making, having clear objectives guides the development and deployment of AI agents.</li> <li><b>Gather and prepare data. </b>AI agents rely on both structured and <a href="https://www.computerweekly.com/feature/How-to-get-structure-from-unstructured-data?">unstructured data to function effectively</a>. Therefore, organizations must ensure they have access to high-quality data that can provide context for the AI agent's tasks. This can include knowledge articles for complex queries and structured data for personalized interactions. Data science experts play a role in obtaining, validating and preparing quality data.</li> <li><b>Select the agent type. </b>Organizations seeking to build an AI agent should choose the most suitable type for their needs. For instance, a reactive agent might suffice for routine customer queries, while more complex tasks requiring adaptability and learning would benefit from a goal-oriented or learning agent that can offer sophisticated support. Suitable ML algorithms and AI operational parameters should be considered early in the agent creation process.</li> <li><b>Consider integration needs with existing systems.</b> When building an AI agent, it's crucial that it seamlessly integrates with existing systems, such as <a href="https://www.techtarget.com/searchcustomerexperience/tip/Understanding-the-3-types-of-CRM-systems">customer relationship management</a> and customer service tools. Such integrations let the AI agent access relevant data and provide better support to users.</li> <li><b>Assemble the development team.</b> The next step is to assemble an AI development team responsible for gathering the data to train the AI agent. The choice of programming languages, platforms and other technologies depends on the team's skills and expertise. ML engineers, data scientists, <a href="https://www.techtarget.com/searchitoperations/definition/DevOps-engineer">DevOps engineers</a>, and user interface and experience designers are a few roles that should be part of the development team. Business leaders should also collaborate with the development team to ensure that the AI agent aligns with business goals and meets the parameters established for strategic business risk.</li> <li><b>Select tools and technologies.</b> It's important to choose the right tech stack. This can include selecting the appropriate technologies, such as ML frameworks; programming languages, including <a href="https://www.techtarget.com/whatis/definition/Python">Python</a> and <a href="https://www.theserverside.com/definition/Java">Java</a>; and AI tools for data processing, model building and training.</li> <li><b>Design the AI agent.</b> The agent's architecture should define how it interacts with users, accesses data and performs various tasks. For more complex agents, this can involve creating a hierarchical structure where higher-level agents manage and direct lower-level agents. Complex tasks involving multiple AI agents will typically involve AI agent orchestration to create cohesive workflows. Extensive model testing and validation is normally part of this phase.</li> <li><b>Train the AI agent.</b> This step involves using the curated data to train the AI agent. It requires feeding data into ML models, enabling the agent to learn patterns, make predictions and refine its decision-making abilities.</li> <li><b>Test and deploy the AI agent. </b>The agent should be thoroughly tested in a controlled environment to assess its performance in various scenarios. Iterative testing helps identify and address issues. Once fully trained and tested, the agent can be deployed in its intended environment, such as local data center or public cloud infrastructure. Early deployments use the AI agent as an option or limit use to select user groups until the agent can render explainable and reliable outcomes. It is important to consider the agent's scalability within the intended infrastructure to ensure that it can continue to perform as user and production data volumes increase.</li> <li><b>Monitor and improve the agent. </b>Lastly, it's crucial to continuously monitor the AI agent's performance, gather feedback and analyze its outcomes. This data should be used to make improvements and updates, enforce data access and security controls, and ensure the agent adapts to changes in user behavior and the business environment.</li> </ol> </section> <section class="section main-article-chapter" data-menu-title="AI agent vendors"> <h2 class="section-title"><i class="icon" data-icon="1"></i>AI agent vendors</h2> <p>Numerous vendor platforms and tools are available for building AI agents. The following is a sample of what's available:</p> <ul class="default-list"> <li><b>AgentGPT. </b>With <a href="https://www.techtarget.com/searchenterpriseai/definition/AgentGPT">AgentGPT</a> users can create, configure and deploy autonomous AI agents in their browser without requiring extensive programming knowledge. Built on OpenAI's GPT-3.5 and <a href="https://www.techtarget.com/whatis/definition/GPT-4">GPT-4</a> models, the platform uses the models' advanced capabilities to generate human-like text and autonomously perform a range of tasks.</li> <li><b>Amazon Bedrock Agents. </b>Amazon Bedrock Agents uses foundation models for reasoning; APIs for communication; and varied data to process user prompts, gather information and complete tasks. Amazon Bedrock Guardrails offers security and multiagent support and collaboration.</li> <li><b>Amazon SageMaker.</b> The <a href="https://www.techtarget.com/searchaws/definition/Amazon-SageMaker">SageMaker</a> fully managed service provides developers and data scientists with tools for building, training and deploying ML models, including AI agents in a production-ready environment. It also offers customizable ML algorithms and infrastructure for scaling.</li> <li><b>Anthropic Claude.</b> Claude is a collection of LLMs developed with a focus on AI ethics. It was developed to be helpful, honest and harmless, and operates as a conversational and multimodal AI that can process text, audio and visual inputs. It's adept at summarizing text, assisting with research, answering questions and writing code.</li> <li><b>Cognition.</b> Cognition builds AI software agents, such as Devin, with the goal of creating AI teammates with advanced reasoning and problem-solving features. These are capable of assisting with coding and optimizing software development tasks.</li> <li><b>Google Cloud Vertex AI. </b>The Vertex AI Agent Builder, which is part of the Google Cloud Vertex AI platform, helps simplify the process of creating autonomous and intelligent agents, enabling both technical and nontechnical users to build them.</li> <li><b>IBM Watson.</b> Watson is IBM's suite of AI tools and applications intended to analyze data, understand language and offer business data insights. It includes IBM watsonx, which provides tools for building and deploying AI models, as well as applications such as Watson Assistant for virtual agent creation.</li> <li><b>LangChain.</b> <a href="https://www.techtarget.com/searchenterpriseai/definition/LangChain">LangChain</a> is a powerful library designed for Python, <a href="https://www.techtarget.com/searchsoftwarequality/news/252529599/Devs-discuss-when-to-use-TypeScript-vs-JavaScript">JavaScript and TypeScript</a> that facilitates the rapid prototyping of LLM-powered applications. It lets developers chain together LLM tasks, which is essential for building complex AI agents.</li> <li><b>Lindy. </b>Lindy is an AI-powered automation platform designed to help businesses create AI agents that collaborate with human teams to automate repetitive tasks. The goal is to free humans from mundane tasks.</li> <li><b>Microsoft AutoGen.</b> The AutoGen open source framework is designed to simplify the process of building and managing AI agents, letting them collaborate and perform tasks autonomously or with human oversight. Through AutoGen, multiple AI agents can work together to solve complex tasks. It uses language models, such as GPT-4, to enhance agent capabilities.</li> <li><b>Microsoft Azure AI.</b> Azure AI is Microsoft's cloud platform offering services and tools for building, deploying and maintaining AI and ML applications. It offers prebuilt and customizable APIs for language processing, vision, speech and decision-making. Additional tools like Azure AI Studio and Azure AI Foundry support the AI development lifecycle.</li> <li><b>OpenAI.</b> OpenAI provides ChatGPT, using the GPT-5 model. It provides a uniform system that operates with a range of AI tasks, such as coding, math, writing, health, and audio and visual perception.</li> <li><b>Salesforce Agentforce. </b>The <a href="https://www.techtarget.com/searchcustomerexperience/news/366610852/Salesforces-ambitions-for-Agentforce-platform-come-to-light">Agentforce platform</a> is designed to create and deploy autonomous AI agents that can support users in business functions, including sales, service and marketing. The platform's low-code Agent Builder helps users define and customize AI agents using natural language queries.</li> <li><b>UiPath.</b> UiPath provides AI agents and enterprise automation software focused on robotic process automation. This incorporates advanced AI and generative AI technologies that support agentic workflow automation. These help companies automate and optimize complex business processes through AI agents.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="Difference between nonagent chatbots, AI assistants, AI agents and generative AI"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Difference between nonagent chatbots, AI assistants, AI agents and generative AI</h2> <p>Nonagent chatbots, AI assistants, AI agents and <a href="https://www.techtarget.com/searchenterpriseai/definition/generative-AI">generative AI</a> are all forms of AI designed to assist users. However, they differ in their capabilities, complexity and real-world applications. The key features and distinctions among these AI technologies include the following:</p> <h3>Nonagent chatbots</h3> <p><b>Functionality.</b> Nonagent chatbots are typically limited to predefined scripts and <a href="https://www.techtarget.com/searchenterpriseai/definition/decision-tree-in-machine-learning">decision trees</a>. They excel at handling simple queries and providing basic information, such as frequently asked questions, but their interactions are often linear and lack the depth and adaptability of AI agents.</p> <p><b>Complexity.</b> Chatbots are simpler to execute than the other AI technologies, but they operate on a predefined set of rules and lack intuitive understanding of human language. They're great for handling straightforward tasks, but they can struggle with complex or unexpected queries.</p> <p><b>User experience. </b>Chatbot interactions can feel rigid because of their scripted nature. This can also lead to less satisfying experiences when users ask questions outside their defined scope.</p> <p><b>Investment costs.</b> Chatbots are easier and cheaper to deploy, making them a popular choice for businesses with limited technical resources.</p> <h3>AI assistants</h3> <p><b>Functionality.</b> AI assistants have more advanced conversational capabilities and context-awareness than chatbots, but they're not as independently functional as autonomous agents.</p> <p><b>Complexity. </b>AI assistants are more complex than nonagent chatbots. They use LLMs that let them interpret nuance and adapt dynamically to unfamiliar or complex requests.</p> <p><b>User experience. </b>AI assistants provide more natural, flexible interactions, often mirroring conversational tone and adjusting responses based on context.</p> <p><b>Investment costs. </b>AI assistants typically require higher upfront investment than chatbots, but their broader capabilities can automate more sophisticated tasks, reducing long-term operational costs.</p> <h3>AI agents</h3> <p><b>Functionality.</b> AI agents are advanced systems capable of autonomously performing and adapting to a range of tasks. They're designed to augment human capabilities and operate across various domains, not just customer service.</p> <p><b>Complexity. </b>Agentic AI systems require more sophisticated technology, including ML and NLP, to understand context and perform tasks effectively. Since they can learn from interactions and improve over time, they're typically suitable for more complex applications.</p> <p><b>User experience.</b> AI agents are conversational systems that deliver a dynamic, engaging user experience by handling multiturn conversations and personalized responses based on user behavior and preferences. They learn from and respond to humans in a natural, human-like way.</p> <p><b>Investment costs.</b> Setting up and running AI agents requires a high initial investment and a skilled team to manage their learning and operational capabilities. This typically includes buying or developing LLMs, acquiring necessary hardware and integrating the system into the existing infrastructure. These systems need large amounts of <a href="https://www.techtarget.com/searchenterpriseai/feature/How-data-quality-shapes-machine-learning-and-AI-outcomes">quality data for training and improving outcomes</a>, additional costs can include data collection, storage and processing.</p> <h3>Generative AI</h3> <p><b>Functionality.</b> Generative AI focuses on generating or synthesizing new information rather than responding to user queries or performing tasks autonomously. This includes generating text, images, music and artwork using models trained on vast data sets.</p> <p><b>Complexity.</b> Generative AI models, such as Open AI <a href="https://www.techtarget.com/whatis/definition/ChatGPT">ChatGPT</a>, often use <a href="https://www.techtarget.com/searchenterpriseai/definition/deep-learning-deep-neural-network">deep learning</a> techniques and large data sets to learn patterns and generate outputs. This requires significant computational resources and sophisticated training processes, making them inherently more complex.</p> <p><b>User experience.</b> Generative AI offers an interactive experience, letting users engage in dynamic conversations that can adapt to their inputs. Users can ask open-ended questions and receive detailed, contextually relevant responses.</p> <p><b>Investment costs.</b> Generative AI requires substantial investment. Training and operating generative AI models, such as those based on LLMs, can <a target="_blank" href="https://www.forbes.com/sites/katharinabuchholz/2024/08/23/the-extreme-cost-of-training-ai-models/" rel="noopener">cost millions</a>. This includes expenses related to data acquisition, computational resources and ongoing maintenance.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/key_attributes_of_ai_agents_vs_nonagent_chatbots_and_generative_ai-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/key_attributes_of_ai_agents_vs_nonagent_chatbots_and_generative_ai-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/key_attributes_of_ai_agents_vs_nonagent_chatbots_and_generative_ai-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/key_attributes_of_ai_agents_vs_nonagent_chatbots_and_generative_ai-f.png 1280w" alt="Table describing attributes of AI agents, nonagent chatbots and generative AI" height="336" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Several key attributes distinguish AI agents from nonagent chatbots and generative AI. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Future of AI agents"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Future of AI agents</h2> <p>A 2025 McKinsey and Company <a target="_blank" href="https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work" rel="noopener">report</a> said that over the next three years, 92% of companies plan to increase their investments in AI. Still, only 1% of businesses have fully integrated AI into workflows where it can drive notable business outcomes. This means AI agents are evolving quickly and are expected to follow their current arc of advancement in three major areas: capability, autonomy and integration.</p> <p>Future AI agents should become even more capable of handling a wide range of tasks. Rather than building more capabilities into larger and more complex AI agents, it's likelier that AI agents will diversify and become more task-specific and specialized within well-defined vertical industries. AI workflow orchestration will tie these agents together to collaborate and perform more complex tasks.</p> <p>AI agents will also become more autonomous as models and data sets become defined and explainability emerges as a key feature of AI systems. This means AI agents will be more robust, able to learn and alter behaviors with greater confidence and success without the need for human intervention or feedback.</p> <p>Finally, future agents will provide greater levels of integration, allowing them to interoperate with a range of business applications, tools, systems, infrastructure and other agents. This will require some standardization around interfaces and protocols, and could pose challenges for legacy systems. However, the demand for AI capabilities is likely to far outweigh the disruption of legacy upgrades.</p> </section> AI agents are autonomous intelligent software components that form the foundation of artificial intelligence (AI) systems. https://cdn.ttgtmedia.com/visuals/digdeeper/2.jpg https://www.techtarget.com/searchenterpriseai/definition/AI-agents Thu, 20 Nov 2025 10:15:00 GMT <p>A neural network is a machine learning (<a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-ML">ML</a>) model designed to process data in a way that mimics the function and structure of the human brain. Neural networks are intricate networks of interconnected nodes, or <a href="https://www.techtarget.com/searchcio/definition/artificial-neuron">artificial neurons</a>, that collaborate to tackle complicated problems.</p> <p>Also referred to as <i>artificial neural networks (ANNs)</i>, neural nets or deep neural networks, neural networks represent a type of <a href="https://www.techtarget.com/searchenterpriseai/definition/deep-learning-deep-neural-network">deep learning</a> technology that's classified under the broader field of <a href="https://www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence">AI</a>.</p> <p>Neural networks are widely used in a variety of applications, including <a href="https://www.techtarget.com/searchenterpriseai/definition/image-recognition">image recognition</a>, predictive modeling, decision-making and natural language processing (<a href="https://www.techtarget.com/searchenterpriseai/definition/natural-language-processing-NLP">NLP</a>). Examples of significant commercial applications over the past 25 years include handwriting recognition for check processing, <a href="https://www.techtarget.com/searchenterpriseai/feature/Speech-to-text-for-deaf-users-aids-in-accessibility">speech-to-text transcription</a>, oil exploration data analysis, weather prediction and <a href="https://www.techtarget.com/searchenterpriseai/definition/facial-recognition">facial recognition</a>.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/yOf2ssqJFFk?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <section class="section main-article-chapter" data-menu-title="How do artificial neural networks work?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How do artificial neural networks work?</h2> <p>An ANN usually involves many <a href="https://www.techtarget.com/whatis/definition/processor">processors</a> operating in parallel and arranged in tiers or layers. These tiers, or layers, fall under three categories -- an input layer, a number of hidden layers, and<i> </i>an output layer. The first tier -- analogous to optic nerves in human visual processing -- receives the raw input information. Its job is to process, analyze, and categorize the incoming data and then pass it on to the next layer.</p> <p>Instead of the original raw input, each successive tier receives the output from the preceding tier, the same way neurons further from the optic nerve receive signals from those closer to it. There may be many hidden layers in an ANN and they all function in the same way -- analyze and process the output from the previous layer and then pass it on to the next layer for further analysis and processing.</p> <p>Simple neural networks have fewer hidden layers. Deep neural networks have several hidden layers and millions of interlinked neurons that may have more or less influence on the other neurons. The large number of layers and neurons allow deep ANNs to process complex problems and map any input type to any output type.</p> <p>The last tier of an ANN, intuitively named the <i>output layer</i>, produces the system's result. This layer may have one or multiple nodes, depending on the problem being addressed by the ANN. For example, <a href="https://www.techtarget.com/whatis/definition/binary">binary</a> classification problems require only one node in the output layer, while multi-class classification problems require multiple output nodes.</p> <p>Each processing node in the ANN has its own small sphere of knowledge, including what it has seen and any rules it was originally programmed with or developed for itself. The tiers are highly interconnected, which means each node in Tier <i>N</i> will be connected to many nodes in Tier <i>N-1</i> -- its inputs -- and in Tier <i>N+1</i>, which provides input data for the Tier <i>N-1</i> nodes. There could be one or more nodes in the output layer, from which the answer it produces can be read.</p> <p>Each individual node or neuron carries information, and the connections between neurons are regulated by weights and biases (also known as thresholds). Weights are assigned to every input layer and they determine how much each variable contributes to the output. A variable with a higher weight contributes more to the output compared to variables with a lower weight.</p> <p>Once the input goes through all the layers of the ANN, the output is determined. First, the inputs are multiplied by their respective weights, then they are summed. This total goes through an activation function, which determines the output. The activation function is important because it enables the neural network to learn more complex patterns over time.</p> <p>Next, the output is compared to the threshold. If the output exceeds the threshold, the node is activated, and the output becomes the input for the next layer in the ANN and is passed to it for further processing. This is known as a feedforward mechanism, meaning information flows only in one direction -- from input to output.</p> <p>Most ANNs are <a href="https://www.techtarget.com/searchenterpriseai/feature/How-neural-network-training-methods-are-modeled-after-the-human-brain">feedforward</a>, although they can also be trained to move in the opposite direction (from output to input). This mechanism, known as backpropagation, is generally used to calculate the error associated with each neuron in the ANN and accordingly adjust the model's parameters.</p> <p>ANN <a href="https://www.techtarget.com/whatis/definition/algorithm">algorithms</a> continuously adjust their weights and bias using <a href="https://www.techtarget.com/searchenterpriseai/definition/reinforcement-learning">reinforcement learning</a> and a method called <a href="https://www.techtarget.com/searchenterpriseai/definition/gradient-descent">gradient descent</a>. The algorithm's goal is to adjust its weights in order to reduce output errors. The more the algorithm is trained, the more its parameters adjust to further reduce errors (also known as minimizing the cost function).</p> <p>ANNs are noted for being <a href="https://www.techtarget.com/searchenterpriseai/tip/Explore-real-world-use-cases-for-adaptive-AI">adaptive</a>, which means they modify themselves as they learn from initial training, and subsequent runs provide more information about the world. The most basic learning model is centered on weighting the input streams, which is how each node measures the importance of input data from each of its predecessors. Inputs that contribute to getting the right answers are weighted higher.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineImages/deep_neural_network.jpg"> <img data-src="https://www.techtarget.com/rms/onlineImages/deep_neural_network_mobile.jpg" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineImages/deep_neural_network_mobile.jpg 960w,https://www.techtarget.com/rms/onlineImages/deep_neural_network.jpg 1280w" alt="A graphic illustrating how the multiple layers in a neural network connect with all the other layers leading to a result" height="342" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Each layer in a neural network consists of small, individual neurons, interlinked to the other neurons. A number called the weight represents the connections between the nodes. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Applications of artificial neural networks"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Applications of artificial neural networks</h2> <p>Image recognition was one of the first areas in which neural networks were successfully applied. A specific type of ANNs called convolutional neural networks (<a href="https://www.techtarget.com/searchenterpriseai/definition/convolutional-neural-network">CNNs</a>) is used for image-related tasks, such as image recognition, pattern recognition and computer vision. CNNs include multiple hidden layers that perform mathematical functions -- specifically, functions from <a href="https://aibusiness.com/ml/five-ways-to-learn-about-machine-learning-online">linear algebra</a> -- to identify patterns and extract relevant features from input images. Different layers extract different features from the input. At the output end, the CNN recognizes the image and can even classify it as a specific type.</p> <p><a href="https://www.techtarget.com/searchenterpriseai/definition/machine-vision-computer-vision">Computer vision technology</a> is another useful application of ANNs. It allows ANNs to identify, extract information from, and classify both images and videos. A deep neural network that's been trained on large volumes of relevant data can perform computer vision tasks at almost the same accuracy -- and much higher speeds -- than humans. Some of the applications of computer vision include the following:</p> <ul class="default-list"> <li>Self-driving cars can recognize road signs, obstacles and people to adjust movement (turn, stop, swerve, etc.).</li> <li>Cameras with facial recognition capabilities can identify human faces and recognize specific attributes, like facial hair, to identify specific individuals.</li> <li>Traffic cameras can detect and flag traffic violations and intelligently manage and optimize traffic flows.</li> <li><a href="https://www.techtarget.com/whatis/definition/medical-imaging">Medical imaging</a> machines can analyze imaging documents to capture useful insights that support diagnostic decision-making, identify tumors, monitor patient vital signs, and track patients' chronic conditions.</li> <li>Robots can identify defects in products on the assembly line or monitor equipment to flag potential issues (predictive maintenance) before they occur.</li> <li>Image labeling systems in retail can capture image details to help retailers pinpoint missing items and to provide precise search results to customers.</li> </ul> <p>Apart from image and video recognition, ANNs are also used for <a href="https://www.techtarget.com/searchcustomerexperience/definition/speech-recognition">speech recognition</a> and NLP.</p> <p>The most obvious example of ANNs recognizing and responding to human speech is their use in <a href="https://www.techtarget.com/searchcustomerexperience/definition/virtual-assistant-AI-assistant">virtual assistants</a>, like Apple's Siri and Amazon's Alexa. Siri and Alexa use ANNs to understand human input and, in response, perform various tasks -- play a song, send an email, show the weather, etc.</p> <p>ANNs also enable computers to understand natural human language and respond in kind. NLP powered by ANNs is used in chatbots, to analyze and summarize documents containing unstructured (e.g., text or images) data, to generate new content (e.g., for marketing), and to perform customer sentiment analysis by analyzing customer content on social media and other places.</p> <p><a href="https://www.techtarget.com/whatis/definition/recommendation-engine">Recommendation engines</a> also rely on ANNs to analyze a user's behavioral and preference history and, accordingly, provide personalized recommendations. Netflix and Amazon are two of the best examples of ANNs driving recommendation engines and enabling the brands to connect with customers in more personalized and scalable ways.</p> <p>In general, the technology <a href="https://www.techtarget.com/searchenterpriseai/ehandbook/Neural-network-applications-in-business-run-wide-fast-and-deep">uses of neural networks</a> have expanded from just image recognition to many additional areas, including the following:</p> <ul class="default-list"> <li><a href="https://www.techtarget.com/searchcustomerexperience/definition/chatbot">Chatbots</a>.</li> <li>Computer vision.</li> <li>NLP, translation and language generation.</li> <li>Speech recognition.</li> <li>Recommendation engines.</li> <li>Stock market forecasting.</li> <li>Delivery driver route planning and optimization.</li> <li>Medical diagnosis and disease recognition.</li> <li>Drug discovery and development.</li> <li><a href="https://www.techtarget.com/whatis/definition/social-media">Social media</a>.</li> <li>Personal assistants.</li> <li><a href="https://www.techtarget.com/whatis/definition/pattern-recognition">Pattern recognition</a>.</li> <li>Sequence recognition.</li> <li>Data processing.</li> <li>Data mining</li> <li>Regression analysis.</li> <li>Process and quality control.</li> <li>Targeted marketing through social network filtering and behavioral data insights.</li> <li>Generative AI.</li> <li><a href="https://www.techtarget.com/whatis/definition/quantum-theory">Quantum chemistry</a>.</li> <li><a href="https://www.techtarget.com/searchbusinessanalytics/definition/data-visualization">Data visualization</a>.</li> <li>Email spam filtering.</li> <li>Financial modeling.</li> <li><a href="https://www.techtarget.com/whatis/definition/robotics">Robotics</a>.</li> <li>Infrastructure reliability analysis.</li> <li>Black-box modeling in geoscience.</li> <li><a href="https://www.techtarget.com/searchsecurity/definition/cybersecurity">Cybersecurity</a>.</li> <li>Financial fraud detection.</li> <li>Materials science research.</li> </ul> <p>The prime uses of ANNs involve any process that operates according to strict rules or patterns and has large amounts of data. If the amount of data involved is too large for a human to make sense of in a reasonable amount of time, the process is likely a good candidate for automation through artificial neural networks.</p> </section> <section class="section main-article-chapter" data-menu-title="How do neural networks learn? How are neural networks trained?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How do neural networks learn? How are neural networks trained?</h2> <p><i>Training</i> a neural network means teaching it how to perform a certain task. Typically, an ANN is initially trained or fed large amounts of data. Training consists of providing input and telling the network what the output should be. For example, to build a network that identifies the faces of actors, the initial training might be a series of pictures, including actors, non-actors, masks, statues and animal faces. Each input is accompanied by matching identification, such as actors' names or "not actor" or "not human" information. Providing the answers enables the model to adjust its internal weighting to do its job better.</p> <p>For example, if nodes David, Dianne and Dakota tell node Ernie that the current input image is a picture of Brad Pitt, but node Durango says it's George Clooney, and the training program confirms it's Pitt, Ernie decreases the weight it assigns to Durango's input and increases the weight it gives to David, Dianne and Dakota.</p> <p>Basically, the ANN first processes a large <a href="https://www.techtarget.com/whatis/definition/data-set">data set</a> that might contain labeled or unlabeled data. This allows it to learn how to process new, previously unseen data. The more data it is trained on, the better its learning capabilities and the more accurate its output over time.</p> <p>In defining the rules and making determinations -- the decisions of each node on what to send to the next layer based on inputs from the previous tier -- neural networks use several principles. These include gradient-based training, <a href="https://www.techtarget.com/searchenterpriseai/definition/fuzzy-logic">fuzzy logic</a>, genetic algorithms and <a href="https://www.techtarget.com/searchenterpriseai/definition/What-is-Bayes-theorem">Bayesian</a> methods. They might be given some basic rules about object relationships in the data being modeled.</p> <p>For example, a facial recognition system might be instructed, "Eyebrows are found above eyes," or "Mustaches are below a nose. Mustaches are above and/or beside a mouth." Preloading rules can make training faster and the model more powerful faster. But it also includes assumptions about the nature of the problem, which could prove to be either irrelevant and unhelpful, or incorrect and counterproductive, making the decision about what, if any, rules to build unimportant.</p> <p>Further, the assumptions people make when training algorithms cause neural networks to amplify cultural biases. <a href="https://www.techtarget.com/searchenterpriseai/feature/6-ways-to-reduce-different-types-of-bias-in-machine-learning">Biased data sets are an ongoing challenge</a> in training systems that find answers on their own through pattern recognition in data. If the data feeding the algorithm isn't neutral -- and almost no data is -- the machine propagates bias.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineImages/bi_ezine-how_ai_systems_amplify_bias.png"> <img data-src="https://www.techtarget.com/rms/onlineImages/bi_ezine-how_ai_systems_amplify_bias_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineImages/bi_ezine-how_ai_systems_amplify_bias_mobile.png 960w,https://www.techtarget.com/rms/onlineImages/bi_ezine-how_ai_systems_amplify_bias.png 1280w" alt="Infographic on how AI amplifies bias" height="336" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>The problem with biased data sets exists in the training of neural systems. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p></p> </section> <section class="section main-article-chapter" data-menu-title="Types of neural networks"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Types of neural networks</h2> <p>Neural networks are sometimes described in terms of their depth, including how many layers they have between input and output, or the model's so-called hidden layers. This is why the term neural network is used almost synonymously with deep learning. Neural networks can also be described by the number of hidden nodes the model has, or in terms of how many input layers and output layers each node has. Variations on the classic neural network design enable various forms of forward and backward propagation of information among tiers.</p> <p>Specific types of ANNs include the following:</p> <h3>Feed-forward neural networks</h3> <p>One of the simplest variants of neural networks, these pass information in one direction, through various input nodes, until it makes it to the output node. The network might or might not have hidden node layers, making their functioning more interpretable. It's prepared to process large amounts of noise. This type of ANN computational model is used in technologies such as facial recognition and computer vision.</p> <h3>Recurrent neural networks</h3> <p>More complex in nature, recurrent neural networks (<a href="https://www.techtarget.com/searchenterpriseai/definition/recurrent-neural-networks">RNNs</a>) save the output of processing nodes and feed the result back into the model. This is how the model learns to predict the outcome of a layer. Each node in the RNN model acts as a memory cell, continuing the computation and execution of operations.</p> <p>This neural network starts with the same front propagation as a feed-forward network, but then goes on to remember all processed information to reuse it in the future. If the network's prediction is incorrect, then the system self-learns and continues working toward the correct prediction during <a href="https://www.techtarget.com/searchenterpriseai/definition/backpropagation-algorithm">backpropagation</a>. This type of ANN is frequently used in text-to-speech conversions.</p> <h3>Convolutional neural networks</h3> <p>Convolutional neural networks (CNNs) are one of the most popular models used today. This computational model uses a variation of multilayer <a href="https://www.techtarget.com/whatis/definition/perceptron">perceptrons</a> and contains one or more convolutional layers that can be either entirely connected or pooled. These convolutional layers create feature maps that record a region of the image that's ultimately broken into rectangles and sent out for nonlinear processing.</p> <p>The CNN model is particularly popular in the realm of image recognition. It has been used in many of the most advanced applications of AI, including facial recognition, text digitization and NLP. Other use cases include paraphrase detection, signal processing and image classification.</p> <h3>Deconvolutional neural networks</h3> <p><a href="https://www.techtarget.com/searchenterpriseai/definition/deconvolutional-networks-deconvolutional-neural-networks">Deconvolutional neural networks</a> use a reversed CNN learning process. They try to find lost features or signals that might have originally been considered unimportant to the CNN system's task. This network model can be used in image synthesis and analysis.</p> <h3>Modular neural networks</h3> <p>These contain multiple neural networks working separately from one another. The networks don't communicate or interfere with each other's activities during the computation process. Consequently, complex or big computational processes can be performed more efficiently.</p> <h3>Perceptron neural networks</h3> <p>These represent the most basic form of neural networks and were introduced in 1958 by Frank Rosenblatt, an American psychologist, who is also considered the father of deep learning. In fact, the perceptron is the oldest neural network.</p> <p>Rosenblatt published his research about perceptrons in the 1958 paper "<a target="_blank" href="https://www.academia.edu/60542953/The_perceptron_a_probabilistic_model_for_information_storage_and_organization_in_the_brain" rel="noopener">The Perceptron: a probabilistic model for information storage and organization in the brain</a>." In the paper, he explained how he got the IBM 704 computer to learn how to distinguish between two sets of cards. Rosenblatt concluded the paper by suggesting that the study of perceptrons might lead to a better understanding of "those fundamental laws of organization which are common to all information handling systems."</p> <p>A single-layer perceptron can only perform simple, linear computational tasks since it only has one layer of neurons between the input and output layers. This makes it unsuitable for more complex tasks.</p> <p>The perceptron takes in the input, weighs it and then sums up the weights and finally produces the output. As with other ANNs, the weights and thresholds of the neurons in a perceptron are adjustable.</p> <p>The perceptron is specifically designed for binary classification tasks, enabling it to differentiate between two classes based on input data. Simple single-layer perceptrons can be built using <a href="https://www.techtarget.com/whatis/definition/open-source">open source</a> machine learning frameworks like TensorFlow<b>.</b></p> <p>Neural networks are not the same as machine learning. In fact, ANNs are a subset of the broader field of machine learning.</p> <h3>Multilayer perceptron networks</h3> <p>Multilayer perceptron (MLP) networks, also known as feedforward neural networks, consist of multiple layers of neurons, including an input layer, one or more hidden layers, and an output layer. Each layer is fully connected to the next, meaning that every neuron in one layer is connected to every neuron in the subsequent layer. Every layer transforms the received input and passes it on to the next layer until the final output is generated at the output layer. This highly interconnected neuronal architecture enables MLPs to learn complex patterns and relationships in data, making them suitable for various classification and <a href="https://www.techtarget.com/searchenterpriseai/feature/What-is-regression-in-machine-learning">regression tasks</a>. MLPs can also be trained to work on <a href="https://www.techtarget.com/searchdatacenter/definition/parallel-processing">parallel computing</a> tasks.</p> <p>Despite the use of the word "perceptron" in the name multilayer perceptron network, MLPs are not comprised of perceptrons but of sigmoid neurons. Also, it uses non-linear activation functions. It is these features that allow MLPs to work on complex, non-linear real-world problems, including computer vision and NLP.</p> <p>One drawback of MLPs is that they are computationally expensive. Since they comprise many layers, training MLPs is a slow process. Also, MLPs are prone to overfitting, which can lead to sub-optimal generalization of new, unseen data.</p> <h3>Radial basis function neural networks</h3> <p>Radial basis function (RBF) neural networks are a type of feed-forward neural networks that use a three-layer architecture. They also have universal approximation capabilities and use radial basis functions, such as <a href="https://www.techtarget.com/whatis/feature/Model-collapse-explained-How-synthetic-training-data-breaks-AI">Gaussian</a> functions, as activation functions. RBFs are typically used for non-linear function approximation and <a href="https://www.techtarget.com/searchenterpriseai/news/365532836/How-a-time-series-forecasting-vendor-uses-Lightning-PyTorch">time series prediction</a> tasks, as well as in control systems. They can learn quickly and offer efficient performance for non-linear system identification, classification and regression problems. Also, the simple architecture makes it easy to understand and implement RBF neural networks -- despite the fact that they require three-stage training.</p> <h3>Transformer neural networks</h3> <p>As one of the more cutting-edge types of neural networks, <a href="https://www.techtarget.com/searchenterpriseai/feature/Transformer-neural-networks-are-shaking-up-AI">transformer neural networks are reshaping NLP</a> and other fields through a range of advancements. Introduced by researchers from Google and the University of Toronto in a <a target="_blank" href="https://arxiv.org/pdf/1706.03762" rel="noopener">2017 paper</a>, transformers are specifically designed to process sequential data, such as text, by effectively capturing relationships and dependencies between elements in the sequence, regardless of their distance from one another.</p> <p>One of the Google researchers also published a follow-up article in which he stated that the transformer is "a novel neural network architecture based on a self-attention mechanism that we believe to be particularly well suited for language understanding." They also explained how they enabled a transformer to outperform both CNNs and RNNs on academic English to German and English to French translation benchmarks.</p> <p>Transformer neural networks have gained popularity as an alternative to CNNs and RNNs because their "attention mechanism" enables them to capture and process multiple elements in a sequence simultaneously, which is a distinct advantage over other neural network architectures. It is this mechanism that makes transformers highly suitable for complex NLP tasks, such as <a href="https://www.smartcitiesdive.com/news/american-city-county-and-smart-cities-dive-are-combining/754194/">text translations</a>.</p> <h3>Generative adversarial networks</h3> <p><a href="https://www.techtarget.com/searchenterpriseai/definition/generative-adversarial-network-GAN">Generative adversarial networks</a> consist of two neural networks -- a generator and a discriminator -- that compete against each other. The generator creates fake data, while the discriminator evaluates its authenticity. These types of neural networks are widely used for generating realistic images and data augmentation processes.</p> </section> <section class="section main-article-chapter" data-menu-title="Advantages of artificial neural networks"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Advantages of artificial neural networks</h2> <p>Artificial neural networks offer the following benefits:</p> <ul class="default-list"> <li><b>Parallel processing.</b> ANNs' parallel processing abilities mean the network can perform more than one job at a time.</li> <li><b>Feature extraction.</b> Neural networks can automatically learn and extract relevant features from raw data, which simplifies the modeling process. However, traditional ML methods differ from neural networks in the sense that they often require manual feature engineering.</li> <li><b>Information storage.</b> ANNs store information on the entire network, not just in a database. This ensures that even if a small amount of data disappears from one location, the entire network continues to operate.</li> <li><b>Nonlinearity.</b> The ability to learn and model nonlinear, complex relationships helps model the real-world relationships between input and output.</li> <li><b>Fault tolerance.</b> ANNs come with <a href="https://www.techtarget.com/searchdisasterrecovery/definition/fault-tolerant">fault tolerance</a>, which means the corruption or fault of one or more cells of the ANN won't stop the generation of output.</li> <li><b>Gradual corruption.</b> This means the network slowly degrades over time instead of degrading instantly when a problem occurs.</li> <li><b>Unrestricted input variables.</b> No restrictions are placed on the input variables, such as how they should be distributed.</li> <li><b>Observation-based decisions.</b> ML means the ANN can learn from events and make decisions based on the observations.</li> <li><b>Unorganized data processing.</b> ANNs are exceptionally good at organizing large amounts of data by processing, sorting and categorizing it.</li> <li><b>Ability to learn hidden relationships.</b> ANNs can learn the hidden relationships in data without commanding any fixed relationship. This means ANNs can better model highly volatile data and nonconstant variance.</li> <li><b>Ability to generalize data.</b> The ability to generalize and infer unseen relationships on unseen data means ANNs can predict the output of unseen data.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="Disadvantages of artificial neural networks"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Disadvantages of artificial neural networks</h2> <p>Their numerous benefits notwithstanding, it's important to note that neural networks also have some drawbacks, including the following:</p> <ul class="default-list"> <li><b>Lack of rules.</b> The lack of rules for determining the proper network structure means the appropriate ANN architecture can only be found through trial, error and experience.</li> <li><b>Computationally expensive.</b> Neural networks such as ANNs use many computational resources. Therefore, training neural networks can be expensive and time-consuming, requiring significant processing power and memory. This can be a barrier for organizations with limited resources or those needing real-time processing.</li> <li><b>Hardware dependency.</b> The requirement of processors with parallel processing abilities makes neural networks dependent on hardware.</li> <li><b>Numerical translation.</b> The network works with numerical information, meaning all problems must be translated into numerical values before they can be presented to the ANN.</li> <li><b>Lack of trust.</b> The lack of explanation behind probing solutions is one of the biggest disadvantages of ANNs. The inability to explain the <i>why</i> or <i>how</i> behind the solution generates a lack of trust in the network.</li> <li><b>Inaccurate results.</b> If not trained properly, ANNs can produce incomplete or inaccurate results.</li> <li><b>Black box nature.</b> Because of their <a href="https://www.techtarget.com/whatis/definition/black-box-AI">black box AI</a> model, it can be challenging to grasp how neural networks make their predictions or categorize data.</li> <li><b>Overfitting.</b> Neural networks are susceptible to <a href="https://www.techtarget.com/whatis/definition/overfitting-in-machine-learning">overfitting</a>, particularly when trained on small data sets. They can end up learning the noise -- focusing on non-relevant factors such as the typeface in a document -- in the training data instead of the underlying patterns, which can result in poor performance on new and unseen data.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="History and timeline of neural networks"> <h2 class="section-title"><i class="icon" data-icon="1"></i>History and timeline of neural networks</h2> <p>The <a href="https://www.techtarget.com/searchenterpriseai/tip/History-of-generative-AI-innovations-spans-9-decades">history of neural networks</a> spans several decades and has seen considerable advancements. The following examines the important milestones and developments in the history of neural networks:</p> <ul class="default-list"> <li><b>1940s.</b> In 1943, mathematicians Warren McCulloch and Walter Pitts built a circuitry system that ran simple algorithms and was intended to approximate the functioning of the human brain.</li> <li><b>1950s.</b> In 1958, Rosenblatt created the perceptron, a form of artificial neural network capable of learning and making judgments by modifying its weights. The perceptron featured a single layer of computing units and could handle problems that were linearly separate.</li> <li><b>1970s.</b> Paul Werbos, an American scientist, developed the backpropagation method, which facilitated the training of multilayer neural networks. It made deep learning possible by enabling weights to be adjusted across the network based on the error calculated at the output layer.</li> <li><b>1980s.</b> Cognitive psychologist and computer scientist Geoffrey Hinton, computer scientist Yann LeCun and a group of fellow researchers began investigating the concept of connectionism, which emphasizes the idea that cognitive processes emerge through interconnected networks of simple processing units. This period paved the way for modern neural networks and deep learning models.</li> <li><b>1990s.</b> Jürgen Schmidhuber and Sepp Hochreiter, both computer scientists from Germany, proposed the long short-term memory recurrent neural network framework in 1997.</li> <li><b>2000s.</b> Hinton and his colleagues at the University of Toronto pioneered restricted Boltzmann machines, a sort of generative artificial neural network that enables <a href="https://www.techtarget.com/searchenterpriseai/definition/unsupervised-learning">unsupervised learning</a>. RBMs opened the path for deep belief networks and deep learning algorithms.</li> <li><b>2010s.</b> Research in neural networks picked up great speed around 2010. The <a href="https://www.techtarget.com/searchdatamanagement/definition/big-data">big data</a> trend, where companies amassed vast troves of data, and parallel computing gave <a href="https://www.techtarget.com/searchenterpriseai/definition/data-scientist">data scientists</a> the training data and computing resources needed to run complex ANNs. In 2011, Google's speech recognition team discovered that deep learning is a powerful approach for speech recognition. This discovery led to large-scale modifications in Google's Android operating system. In the same year, a collaboration between machine learning researcher Andrew Ng and Googler Jeff Dean led to the development of a massively large unsupervised neural network that offered unprecedented performance on computer vision tasks. In 2012, a neural network named AlexNet won the ImageNet Large Scale Visual Recognition Challenge, an image classification competition. Around this time, researchers at Caltech and elsewhere discovered that graphics processing units (<a href="https://www.techtarget.com/searchvirtualdesktop/definition/GPU-graphics-processing-unit">GPUs</a>) could be used to handle the huge computational demands of large/complex neural networks.</li> <li><b>2020s and beyond.</b> Neural networks continue to undergo rapid development, with advancements in architecture, training methods and applications. Researchers are exploring new network structures such as transformers and <a href="https://www.techtarget.com/searchenterpriseai/definition/graph-neural-networks-GNNs">graph neural networks</a>, which excel in NLP and understanding complex relationships. Similarly, Kolmogorov-Arnold Networks (KANs) are being studied for applications involving non-linear and interdependent variable relationships, such as weather modeling and fluid dynamics. Additionally, techniques such as transfer learning and self-supervised learning are enabling neural networks and deep learning models to learn from smaller data sets and generalize better. These developments are driving progress in fields such as healthcare, autonomous vehicles, facial recognition, language translations, and wearables.</li> </ul> <p><i>Discover the process for building a machine learning model, including data collection, preparation, training, evaluation and iteration. Follow these </i><a href="https://www.techtarget.com/searchenterpriseai/feature/How-to-build-a-machine-learning-model-in-7-steps"><i>essential steps to kick-start your ML project</i></a><i>.</i></p> </section> A neural network is a machine learning (ML) model designed to process data in a way that mimics the function and structure of the human brain. https://cdn.ttgtmedia.com/visuals/digdeeper/5.jpg https://www.techtarget.com/searchenterpriseai/definition/neural-network Mon, 27 Oct 2025 12:00:00 GMT <p><strong><a href="#A-C">A-C</a></strong>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <strong><a href="#D-F">D-F</a></strong>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <strong><a href="#G-I">G-I</a></strong>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <strong><a href="#J-L">J-L</a></strong>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <strong><a href="#M-O">M-O</a></strong>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <strong><a href="#P-R">P-R</a></strong>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <strong><a href="#S-U">S-U</a></strong>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <strong><a href="#V-X">V-X</a></strong>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <strong><a href="#Y-Z">Y-Z</a></strong></p> <p>Ensuring the health of people, the planet and even businesses rests on more sustainable practices, which depend on understanding core concepts and terms.</p> <p>Many individuals, businesses, nonprofits and government entities are working hard to quickly create and execute specific sustainability strategies and environmental, social and governance (<a href="https://www.techtarget.com/whatis/definition/environmental-social-and-governance-ESG">ESG</a>) initiatives. Everyone has a role to play. Learning about sustainability and ESG concepts ensures that critical stakeholders have productive conversations that avoid term misuse and oversimplification.</p> <p>Climate disruption is having and will continue to have a negative impact on life across the globe. It's imperative that business leaders, private citizens and governments take immediate and decisive <a target="_blank" href="https://www.ipcc.ch/report/ar6/syr/resources/spm-headline-statements" rel="noopener">action</a>.</p> <p>Here are some fundamental sustainability terms and ESG concepts that provide a foundation for taking action.</p> <section class="section main-article-chapter" data-menu-title="The 50-plus sustainability and ESG terms you should know"> <h2 class="section-title"><i class="icon" data-icon="1"></i>The 50-plus sustainability and ESG terms you should know</h2> <ol class="default-list"> <li><b>business sustainability</b>. Also known as corporate sustainability, <a href="https://www.techtarget.com/whatis/definition/business-sustainability">business sustainability</a> is the ethical and responsible management of an organization's continued success with environmental, social and financial concerns.</li> <li><b>carbon credit</b>. When companies create carbon offsetting initiatives, they receive a transferable or tradeable carbon credit, or token. The credit represents the right to emit greenhouse gases and offset them elsewhere. One credit represents one ton of carbon dioxide reduced or removed from the atmosphere. In practice, taking advantage of these credits lets owners reduce greenhouse gas emissions to get closer to net zero. The term also refers to purchased credits that will fund emission-reducing projects.</li> <li><b>carbon footprint</b>. A <a href="https://www.techtarget.com/whatis/definition/carbon-footprint">carbon footprint</a> measures the amount of carbon dioxide and methane produced by individuals, organizations, products or practices.</li> <li><b>carbon neutral</b>. The ideal balance between carbon dioxide emissions produced by human activity and carbon absorption by the atmosphere; the calculation should come to zero.</li> <li><b>carbon offset</b>. A <a href="https://www.techtarget.com/whatis/definition/carbon-offset">carbon offset</a> is an activity or purchase intended to compensate for carbon emissions produced by individuals and organizations. Carbon storage through tree planting or land restoration is a common example. Businesses with carbon offset programs receive carbon tokens.</li> <li><b>carbon token</b>. A digital asset governed by a <a href="https://www.techtarget.com/searchcio/feature/Examples-of-smart-contracts-on-blockchain">smart contract on blockchain</a> that represents a real-world reduction in one metric ton of carbon dioxide emissions. The asset exists to verify ownership and simplify the carbon credit trading process. Another example is a non-fungible token, or <a href="https://www.techtarget.com/whatis/definition/nonfungible-token-NFT">NFT</a>, representing a single, unique share of captured carbon dioxide associated with a specific time and place. The dependence on blockchain technology to administer carbon tokens is controversial due to blockchain's energy-intensive processes.</li> <li><b>circular economy</b>. The <a href="https://www.techtarget.com/whatis/definition/circular-economy">circular economy</a> keeps products in circulation to the fullest extent possible by reducing material consumption, streamlining processes and collecting waste for reuse.</li> <li><b>clean tech</b>. Sometimes referred to as green technologies or eco-technologies, this term refers to technologies and processes designed to limit negative environmental impacts, such as waste and carbon emissions, especially compared to fossil fuels. Examples of clean technologies include solar power, wind power, biofuels, recycling and smart lighting.</li> </ol> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/Dg8nTJ0cHVI?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <ol start="9" class="default-list"> <li><b>climate adaptation</b>. The act of preparing for and adjusting to climate change's current and projected consequences. For example, cities can build seawalls to protect from rising sea levels.</li> <li><b>climate change</b>. The shifts over time in the average temperature and weather patterns that define specific locations. In particular, climate change has come to mean the rise in global temperatures from heat-trapping gases resulting from mining and using oil, coal and other fossil fuels. <a href="https://www.techtarget.com/sustainability/feature/How-does-climate-change-affect-businesses-Financial-impacts">Climate change indicators</a> include rising sea levels; an increase in the severity of extreme weather, such as hurricanes, droughts and floods; and ice loss at the Earth's poles.</li> <li><b>climate mitigation</b>. The process of decreasing the flow of heat-trapping pollution. For example, reducing fossil fuel burning by using renewable energy sources may help.</li> <li><b>climate resilience</b>. The ability to support a community, company or the natural environment before, during and after a climate event in a timely, efficient manner. Climate resilience differs from climate adaptation, although the two terms are often used interchangeably.</li> <li><b>climate</b><b> risk.</b> As wildfires, droughts, food scarcity, hurricanes and other climate change effects happen, businesses become more vulnerable. Climate risk describes that vulnerability. It is the potential for climate change to create negative effects on human or ecological systems. Risks fall into two main categories: risks based on the transition to a greener economy, such as losing market share by moving away from fossil fuel-based products, and risks related to the physical effects of climate change, such as flooded offices.</li> <li><b>closed-loop</b>. A production process that reuses material waste to create additional products or repurposes recycled materials.</li> <li><b>conscious capitalism</b>. Conscious capitalism is a socially responsible framework for capitalism in the corporate and political spheres. It emphasizes creating human value alongside profit value.</li> <li><b>corporate social responsibility (CSR)</b>. For-profit companies use the <a href="https://www.techtarget.com/searchcio/definition/corporate-social-responsibility-CSR">CSR</a> business model to gauge social and environmental benefits alongside organizational goals, such as profitability.</li> </ol> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/esg_vs_csr_vs_sustainability-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/esg_vs_csr_vs_sustainability-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/esg_vs_csr_vs_sustainability-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/esg_vs_csr_vs_sustainability-f.png 1280w" alt="ESG vs. CSR vs. sustainability: what's the difference?" height="262" width="560"> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <ol start="17" class="default-list"> <li><b>digital carbon footprint</b>. The digital carbon footprint is the amount of greenhouse gas emissions digital devices, tools and platforms produce. All tech, from cloud computing to mobile phones to internet usage, <a href="https://www.techtarget.com/sustainability/tip/Ways-to-reduce-an-organizations-digital-carbon-footprint">produces a digital carbon footprint</a>.</li> <li><b>digital sobriety</b>. Digital sobriety aims to limit the harmful environmental impact of smartphones, internet usage, digital media and other tech in both large and small ways on a daily basis. Moving toward digital sobriety includes a wide range of actions: buying fewer devices, deleting emails, opting for lower-definition media consumption, sustainably developing software and buying less-powerful machines.</li> <li><b>drawdown</b>. A drawdown is the point at which atmospheric greenhouse gas levels stop climbing and start declining.</li> <li><b>electronic waste (e-waste)</b>. Electronics at or nearing the end of their useful life. Green tech and sustainability approaches seek to extend the useful life of devices and use circular economy principles to keep the amount of e-waste to an absolute minimum. The priority is to first reduce waste, then refurbish devices, then move toward recycling.</li> <li><b>energy efficiency</b>. The same task or result is achieved with less energy. For example, heating, cooling and operating appliances and electronics are less energy-intensive in energy-efficient homes and buildings.</li> <li><b>environmental justice</b>. Environmental justice seeks to ensure fair treatment of all people regardless of race, color, national origin or income equally, regarding environmental laws, regulations and policies. The approach holds that no group should bear a disproportionate share of negative environmental consequences.</li> <li><b>environmental, social and governance</b>. Sustainable and ethical interests that can be central to an organization's financial and corporate interests. Otherwise known as ESG.</li> </ol> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/3_pillars_of_esg-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/3_pillars_of_esg-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/3_pillars_of_esg-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/3_pillars_of_esg-f.png 1280w" alt="An explanation of the three pillars of ESG: environmental, social and governance." height="286" width="560"> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p></p> <ol start="24" class="default-list"> <li><b>ESG framework</b>.&nbsp;A set of objectives companies can use in their&nbsp;<a target="_blank" href="https://www.techtarget.com/sustainability/definition/ESG-reporting" rel="noopener">ESG reporting</a>. A variety of&nbsp;<a href="https://www.techtarget.com/sustainability/feature/Top-ESG-reporting-frameworks-explained-and-compared">ESG frameworks</a> exist to help companies evaluate their environmental and social impact and assess their internal governance policies, as well as their risks and opportunities.</li> <li><b>ESG reporting</b>. A disclosure that an organization must provide to show how it works to fulfill its ESG promises and efforts. Some organizations are required to report on ESG, depending on the laws and regulations of the country or countries in which they operate.</li> <li><b>feed-in tariff</b>. A policy designed to accelerate investments in renewable energy. A policy of this type usually involves long-term government contracts.</li> <li><b>global warming</b>. Global warming refers to the heating of the Earth's surface from trapped greenhouse gases resulting from human activities, such as transportation, agriculture, overfishing, fossil fuel energy production and overconsumption. Unless companies, governments and consumers make major shifts, global warming and climate change will heat the planet so much that it will be unlivable in the near future.</li> <li><b>green cloud</b>. The <a href="https://www.techtarget.com/searchstorage/definition/green-cloud">green cloud</a> refers to the possible environmental benefits of IT services delivered over the internet. Typically seen as a buzzword, reliance on the alleged benefits can lead technologists to believe that further efforts to reduce carbon footprints are unnecessary.</li> <li><b>green computing</b>. The sustainable approach to using computing devices and equipment is <a href="https://www.techtarget.com/searchdatacenter/definition/green-computing">green computing</a>. Some methods include reducing resource use, responsible disposal of e-waste and deploying energy-efficient IT equipment.</li> <li><b>green IT</b>. The practice of designing, manufacturing, operating and disposing of IT products and devices to minimize the negative effects of IT operations on the environment is <a href="https://www.techtarget.com/searchcio/definition/green-IT-green-information-technology">green IT</a>.</li> <li><b>green premium</b>. Coined by Bill Gates, green premium refers to the economic and environmental costs of choosing clean tech over financially sound options with higher greenhouse gas emissions.</li> <li><b>green software</b>. <a href="https://www.techtarget.com/searchsoftwarequality/definition/green-software">Green software</a> refers to applications that are designed, developed and implemented in ways intended to minimize energy consumption and environmental effects.</li> <li><b>greenhouse effect</b>. The result of carbon dioxide, methane and nitrous oxides in Earth's atmosphere trapping the sun's heat.</li> <li><b>greenhouse gas emissions</b>. The sum of emissions of various heat-trapping gases. Greenhouse gases include carbon dioxide, methane, nitrous oxides and fluorinated gases such as hydrofluorocarbons.</li> <li><b>Greenhouse Gas Protocol</b>. A globally recognized set of reporting and accounting frameworks for managing greenhouse gas emissions from private and public sector operations, value chains and mitigation actions.</li> <li><b>greenhushing</b>. Greenhushing involves companies intentionally hiding sustainability goals. Companies may do this out of fear of greenwashing accusations or falling short of stated goals.</li> <li><b>greenwashing</b>. Deceptive, misleading or false claims or actions that an organization, product or service has a positive environmental effect is called <a href="https://www.techtarget.com/whatis/definition/greenwashing">greenwashing</a>. Whether intentional or unintentional, the practice is detrimental.</li> </ol> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/beware_of_greenwashing-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/beware_of_greenwashing-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/beware_of_greenwashing-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/beware_of_greenwashing-f.png 1280w" alt="Beware of greenwashing and bluewashing, advice on how to select an ecolabel and understand the social responsibility of an organization's supply chain." height="361" width="560"> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p></p> <ol start="38" class="default-list"> <li><b>high emitters</b>. A designation given to companies or countries that emit comparatively high volumes of greenhouse gas. Per capita emissions are used to measure the emissions of nations.</li> <li><b>impact investing</b>. An investing strategy that directs money toward companies that create a measurable, positive change in the world. This may also be called socially responsible investing.</li> <li><b>impact sourcing.</b> A sourcing strategy that directs employment and career development opportunities toward people from economically disadvantaged backgrounds.</li> <li><b>Intergovernmental Panel on Climate Change (IPCC).</b> The United Nations' body for evaluating scientific climate change information. The IPCC releases regular reports on climate impacts and risk, and offers options for mitigation and adaptation.</li> <li><b>loss and damage</b>. Climate-related consequences that people are unable to adapt to, either because the consequence is too severe or the affected community lacks access to the resources to adapt. Loss and damage result from sudden natural disasters, such as floods, or gradual change, such as desertification.</li> <li><b>materiality assessment</b>. A materiality assessment is a formal way of assessing stakeholders' commitment to specific ESG issues and calculating an organization's <a href="https://www.techtarget.com/sustainability/definition/ESG-score">ESG score</a>. It identifies the effect of a certain issue on a company's performance and competitiveness in the market.</li> <li><b>net zero</b>. The result of lowering greenhouse gas emissions as close to zero as possible and balancing remaining emissions with removals.</li> <li><b>Paris Agreement.</b> The Paris Agreement is a legally binding international treaty on climate change that aims to limit global warming to a 1.5°C temperature increase by the end of the century. The Agreement was adopted at the 2015 UN Climate Change Conference.</li> <li><b>recycling</b>. The process of collecting and processing waste materials, ideally to make new products.</li> <li><b>responsible innovation</b>. Responsible innovation prioritizes ethics and social responsibility in the research, design and production of new technologies or evolutions of existing technology. Responsible innovation posits ethics as a design problem.</li> <li><b>scope 1, 2, 3 emissions</b>. Developed by the Greenhouse Gas Protocol, <a href="https://www.techtarget.com/sustainability/feature/Scope-1-2-and-3-emissions-Differences-with-examples">scopes give organizations a way to categorize</a> their emissions. Organizations may find it easier to control scopes 1 and 2, but scope 3 emissions are the most difficult to track.</li> </ol> <ul class="default-list"> <li style="list-style-type: none;"> <ul class="default-list"> <li><b>scope 1 emissions</b>. The direct emissions generated by an organization's operations. Running machinery, manufacturing products, driving vehicles, heating buildings and providing power to devices generate emissions.</li> <li><b>scope 2 emissions</b>. The indirect emissions generated by an organization's energy purchase and usage. Investment in renewable energy sources may help lower these emissions.</li> <li><b>scope 3 emissions</b>. The indirect emissions generated by an organization's customer and supplier activities.</li> </ul> </li> </ul> <ol start="49" class="default-list"> <li><b>supply chain traceability</b>. In sustainability, <a href="https://www.techtarget.com/searcherp/definition/traceability">traceability</a> not only identifies, tracks and traces materials and commodities, but it also verifies sustainability claims across the value chain.</li> </ol> <div class="extra-info"> <div class="extra-info-inner"> <h3>For more on ESG strategy and management, read the following articles:</h3> <p><a href="https://www.techtarget.com/sustainability/feature/5-ways-organizations-can-address-the-social-factors-of-ESG">The social 'S' in ESG: Examples, factors and best practices</a></p> <p><a href="https://www.techtarget.com/sustainability/feature/A-timeline-and-history-of-ESG-investing-rules-and-practices">A timeline and history of ESG investing, rules and practices</a></p> <p><a href="https://www.techtarget.com/sustainability/feature/ESG-audit-checklist-steps-for-success">ESG audit checklist: Steps for success</a></p> <p><a href="https://www.techtarget.com/sustainability/feature/ESG-marketing-Why-its-important-and-how-to-draft-a-plan">ESG marketing: Why it's important and how to draft a plan</a></p> <p><a href="https://www.techtarget.com/sustainability/feature/ESG-materiality-assessments-What-CIOs-others-need-to-know">ESG materiality assessments: What businesses need to know</a></p> <p><a href="https://www.techtarget.com/sustainability/feature/ESG-metrics-Tips-and-examples-for-measuring-ESG-performance">ESG metrics: Tips and examples for measuring ESG performance</a></p> <p><a href="https://www.techtarget.com/searchcio/feature/10-key-ESG-and-sustainability-trends-for-business-IT">Key ESG and sustainability trends, ideas for companies</a></p> <p><a href="https://www.techtarget.com/sustainability/feature/ESG-vs-CSR-vs-sustainability-Whats-the-difference">ESG vs. CSR vs. sustainability: What's the difference?</a></p> </div> </div> <ol start="50" class="default-list"> <li><b>sustainability</b>. The ability to meet present needs without compromising the needs of future generations. In practice, sustainability aligns environmental protection, human well-being and economic development.<b></b></li> <li><b>triple bottom line (TBL)</b>. According to the <a href="https://www.techtarget.com/whatis/definition/triple-bottom-line-3BL">TBL</a> accounting framework, the bottom lines calculate financial performance alongside environmental and social effects.</li> <li><a id="V-X"></a><a id="Y-Z"></a><b>zero waste</b>. The concept of managing products, packaging and materials responsibly to minimize environmental harm.</li> </ol> <p><b>Editor's note: </b><i>This article was updated to reflect changes in terminology around ESG and sustainability topics.</i></p> <p><i>Guilliean Pacheco is a former associate site editor for TechTarget Editorial's CIO, ERP and Sustainability &amp; ESG sites.</i></p> <p><i>Ben Lutkevich is site editor for TechTarget's IT Infrastructure group. Previously, he wrote definitions and features for Whatis.com.</i></p> </section> Sustainable strategies require a basic understanding of the fundamentals. Business and IT leaders can benefit from this list of essential sustainability terms and ESG concepts. https://cdn.ttgtmedia.com/rms/onlineimages/esg_a468795958.jpg https://www.techtarget.com/sustainability/feature/Sustainability-and-ESG-glossary-Terms-to-know Thu, 09 Oct 2025 00:00:00 GMT <p>Cloud admins and architects need reliable tools to automate launching and managing cloud-based infrastructure. Automation plays a key role in delivering efficient procedures, like code deployments and regular infrastructure updates, to launch modern software applications. Infrastructure as code has become an essential tool to achieve this capability.</p> <p>Users who have an AWS environment and want to employ Infrastructure as code (<a href="https://www.techtarget.com/searchitoperations/definition/Infrastructure-as-Code-IAC">IaC</a>) have two popular options to choose from: AWS CloudFormation and Terraform. CloudFormation is a service native to AWS and Terraform is an open source IaC tool that supports multiple cloud platforms. While they both enable cloud infrastructure deployment automation, they have different approaches to syntax, processes, visibility and resource management.</p> <p>Let's better understand both IaC tools and see how they compare in key features, such as the following:</p> <ul class="default-list"> <li>Modularity.</li> <li>Template customizations.</li> <li>Scalability.</li> <li>Failure handling.</li> <li>Support.</li> <li>DevOps.</li> </ul> <section class="section main-article-chapter" data-menu-title="Choose the right IaC tool"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Choose the right IaC tool</h2> <p>IaC delivers a code-based approach to launching and configuring infrastructure resources such as compute, storage, networking and security. It requires a text-based template using a standardized syntax for each infrastructure resource. This enables application owners to apply version control and minimize manual intervention in the release process. IaC saves software development teams a significant amount of time and delivers consistency and reliability across application environments. Reusable templates and configurations also help with process efficiency across the organization.</p> <h3>What is AWS CloudFormation?</h3> <p>AWS CloudFormation, launched in 2011, is an IaC service that enables users to model and set up AWS resources using templates. The service provisions and manages these resources repeatably and predictably. The learning curve is simple since most developers are already familiar with JSON and YAML syntax. When using CloudFormation, it is highly recommended to use YAML instead of JSON, given that it's easier to handle. It is less verbose than JSON and enables user comments, which is an essential feature for team members to review existing templates. Four important concepts in using AWS CloudFormation are templates, stacks, change sets and stack sets.</p> <h3>What is Terraform?</h3> <p>HashiCorp Terraform, launched in 2014, enables IT teams to automate infrastructure provisioning with reusable, shareable and human-readable configuration files for both on-premises and cloud environments. Terraform has its own template syntax called&nbsp;HashiCorp Configuration Language, more commonly known as HCL. Because HCL is unfamiliar, developers might require some additional time to learn it.</p> <p>In 2023, Hashicorp announced it would adopt a Business Source License, a significant move away from its open-source roots. While Terraform currently holds 62% of the IaC market share, only 47% of practitioners plan to continue to use it in the future, according to "<a href="https://www.firefly.ai/state-of-iac-2025">The State of IaC 2025</a>" by Firefly. OpenTofu, an open source Terraform fork, is rising as a competitor -- 12% of practitioners currently use it and 27% plan to use it in the future.</p> </section> <section class="section main-article-chapter" data-menu-title="Compare key features"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Compare key features</h2> <p>AWS CloudFormation and Terraform use the concept of a stack, which is a grouping of cloud components admins manage as a deployment unit. Stacks can arrange applications, environments and any grouping that is relevant to a particular organization. Parameters define cloud resources specific to the type of cloud components that launch. Both tools use dynamic parameters relevant to a launch, for example, launching different <a href="https://www.techtarget.com/searchcloudcomputing/tip/Use-this-EC2-instance-type-comparison-to-power-your-AWS-apps">Amazon EC2 instance types</a> for a development environment versus a production one.</p> <p>Let's take a closer look at some of the key capabilities of both tools to discover their use cases. We will focus on the following:</p> <h3>Modularity</h3> <p>Modularity is the ability to create common components for reuse across multiple deployments. AWS CloudFormation offers the modules feature, which are building blocks that complement common registries. Administrators can reuse them across multiple stacks within a single account or&nbsp;<a href="https://www.techtarget.com/searchcloudcomputing/tip/What-you-need-to-know-to-manage-multiple-AWS-accounts">across multiple AWS accounts</a>.</p> <p>Terraform has a similar feature, also called modules. It reuses common configurations and manages them either locally or remotely in custom registries or Git repositories.</p> <h3>Template customizations</h3> <p>Terraform provides a wider range of built-in functions for template customization compared to AWS CloudFormation. In many cases, it's useful to apply dynamic configurations to a stack based on conditions and custom logic. Terraform offers more flexibility given the narrow range of AWS CloudFormation built-in functions. However, AWS CloudFormation offers built-in integration with <a href="https://www.techtarget.com/searchcloudcomputing/tip/How-to-create-an-AWS-Lambda-function">custom Lambda functions</a> -- built by the stack owner -- that can perform complex logic during stack updates.</p> <h3>Scalability</h3> <p>CloudFormation and Terraform stacks can manage up to 500 resources, which is sufficient for most large cloud infrastructure deployments. If a particular deployment exceeds this limit, one alternative in either platform is to launch multiple stacks. CloudFormation also offers the nested stack feature, with a limit of 2,500 resources.</p> <p>The amount of time it takes to launch multiple resources can be similar in both tools, but it can vary significantly depending on the type and number of resources users must manage. While some deployments in either platform can take a few seconds to complete, there are situations where they can take several minutes. A key factor in CloudFormation launch time is the interdependency of resources, which results in a sequential launch of these components and potentially a longer launch time.</p> <h3>Failure handling</h3> <p>It's not unusual to face situations where updates aren't applied for various reasons, ranging from incorrect configurations to external failures. AWS CloudFormation delivers a reliable failure handling mechanism that enables developers to either keep, update or roll back resources in an unsuccessful stack creation or update. Terraform doesn't offer a native rollback feature, which means developers must specifically apply the next steps to either remove or update the affected resources.</p> <p>AWS CloudFormation's change sets provide developers with a high-level visualization of resource updates before application. They can then either continue or cancel the operation. Terraform has a similar command called plan, which reduces the likelihood of applying unwanted, potentially destructive updates to a stack.</p> <h3>Support</h3> <p>One important difference between these two IaC tools is that Terraform supports multiple cloud providers, such as AWS, Azure, Google Cloud, Oracle and Digital Ocean, while CloudFormation only supports AWS. For a multi-cloud approach, consider Terraform. Keep in mind that each cloud provider requires a different set of parameters and configurations that users need to specify in Terraform, which could result in a complex set of templates.</p> <p>Given that Terraform is not a cloud provider tool, new features in a particular platform are not necessarily available in Terraform. In the case of AWS, CloudFormation has a shorter time frame before new AWS features and services are available for launch using IaC, compared to Terraform. When using CloudFormation, there's the option to engage AWS support for any potential issues. Terraform doesn't offer direct support from cloud providers, just from Hashicorp and online communities.</p> <h3>DevOps</h3> <p>IaC is an essential component in a DevOps strategy. Integrating cloud infrastructure updates into the code management process delivers critical visibility into how infrastructure deploys, as well as launch and test automation, troubleshooting and rollback strategies in case of failure.</p> <p>Given the potentially disruptive nature of infrastructure updates, it is important to implement manual approval processes for certain infrastructure updates. Terraform's plan command and CloudFormation's change sets are critical components in this process.</p> <p>Both Terraform and CloudFormation deliver a similar level of integration with <a href="https://www.techtarget.com/searchsoftwarequality/CI-CD-pipelines-explained-Everything-you-need-to-know">CI/CD pipelines</a> and DevOps processes. Users can trigger them from pipelines in common Git repositories, such as GitLab, BitBucket or GitHub, or from AWS services such as CodePipeline or CodeBuild.</p> </section> <section class="section main-article-chapter" data-menu-title="How to decide which is right for your organization?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How to decide which is right for your organization?</h2> <p>If there are existing CI/CD pipelines, it is critical to evaluate how each tool would integrate with them and identify any potential advantages or issues. One factor to consider is that Terraform offers a more direct way to implement advanced custom logic within a template.</p> <p>Multi-cloud support vs. AWS-exclusive tech is also a key element to evaluate. In a multi-cloud organization, Terraform will simplify the launch of cloud infrastructure components, while CloudFormation will only support AWS. If an organization's infrastructure launches exclusively in AWS and there are no foreseeable plans to use other cloud platforms, then CloudFormation is likely the best choice. Given that it is a product built by AWS, it will be eligible for AWS Support, and it will likely deliver more useful features for AWS components.</p> <p><i>Editor's note: This article was republished to improve the reader experience and updated to reflect HashiCorp's adoption of a Business Source License.</i></p> <p><i>Ernesto Marquez is owner and project director at Concurrency Labs, where he helps startups launch and grow their applications on AWS. He enjoys building serverless architectures, building data analytics solutions, implementing automation and helping customers cut their AWS costs.</i></p> </section> AWS users need an IaC tool to automate the deployment and management of their AWS environment. AWS CloudFormation and Terraform can both accomplish this goal, but which is best? https://cdn.ttgtmedia.com/rms/onlineimages/cloud_g1213812869.jpg https://www.techtarget.com/searchcloudcomputing/tip/AWS-CloudFormation-vs-Terraform-How-to-choose Thu, 04 Sep 2025 09:00:00 GMT <p>AWS Certified Solutions Architect - Associate is a <a href="https://www.techtarget.com/searchaws/definition/AWS-certification">certification</a> for IT professionals who use <a href="https://www.techtarget.com/searchaws/definition/Amazon-Web-Services">Amazon Web Services</a> to design and implement cloud-based solutions. The certification, which is <a href="https://aws.amazon.com/certification/certified-solutions-architect-associate/" target="_blank" rel="noopener">offered by Amazon</a>, verifies that the certificate holder has successfully demonstrated their technical knowledge and skills across a wide range of AWS tools.</p> <p>This type of certification also validates the certificate holder's understanding of AWS <a href="https://aws.amazon.com/architecture/well-architected/" target="_blank" rel="noopener">Well-Architected Framework</a>. This framework provides <a href="https://www.techtarget.com/searchcloudcomputing/definition/cloud-architect">cloud architects</a> with a structured approach to designing and managing secure, reliable and cost-optimized cloud workloads. AWS developed it to help architects build cloud solutions that are aligned with AWS best practices and make the best use of AWS tools.</p> <p>The AWS Certified Solutions Architect - Associate certification is valid for three years. When cloud architects earn and maintain a Solutions Architect - Associate certification from Amazon, it can increase their earning potential and enhance job security. The certification can also be used as a steppingstone toward more advanced AWS credentials, like <a href="https://aws.amazon.com/certification/certified-solutions-architect-professional/" target="_blank" rel="noopener">AWS Certified Solutions Architect - Professional</a>.</p> <p>Another important benefit of earning an Associate certification for <a href="https://www.techtarget.com/searchcloudcomputing/definition/cloud-architecture">cloud architecture</a> is that it can provide IT professionals with more career options. AWS Associate cloud architect certifications are widely recognized as proof of a candidate's technical expertise and ability to design scalable, secure cloud solutions. Indirectly, this can help candidates get interviews for a number of related cloud-focused IT jobs, including <a href="https://www.techtarget.com/searchcloudcomputing/definition/cloud-engineer">cloud engineer</a>, <a href="https://www.techtarget.com/searchitoperations/definition/DevOps-engineer">DevOps engineer</a> and <a href="https://www.techtarget.com/searchcloudcomputing/definition/cloud-infrastructure">cloud infrastructure</a> engineer.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/IV7wWa-V0-Q?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <section class="section main-article-chapter" data-menu-title="AWS Certified Solutions Architect - Associate: Exam details"> <h2 class="section-title"><i class="icon" data-icon="1"></i>AWS Certified Solutions Architect - Associate: Exam details</h2> <p>AWS Certified Solutions Architect - Associate certification candidates must take Amazon's SAA-C03 exam to earn certification and the badge that comes with it. The exam, which costs $150, requires the applicant to answer 65 questions in 130 minutes.</p> <p>Fifty of the questions are either multiple choice or multiple response. Multiple <i>choice</i> questions have one correct response and three incorrect responses. Multiple <i>response</i> questions have two or more correct responses. All unanswered questions are counted as incorrect answers, so there's no penalty for guessing.</p> <p>The SAA-C03 exam also mixes in 15 questions that do not count in the final score. AWS uses the 15 pilot questions to help plan the next version of the exam. Because pilot questions are not identified, exam takers do not know which questions in the exam count and which do not.</p> <p>To pass the exam and receive certification, exam takers must achieve a scaled score of at least 720 out of 1,000. AWS uses a scaled score (100–1,000) rather than raw percentages to ensure candidates who take different versions of the same exam are assessed fairly.</p> </section> <section class="section main-article-chapter" data-menu-title="How to take the exam"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How to take the exam</h2> <p><a href="https://www.pearsonvue.com/us/en/test-centers.html" target="_blank" rel="noopener">Pearson VUE</a> and <a href="https://www.psiexams.com/" target="_blank" rel="noopener">PSI</a> are AWS' official testing partners. Pearson VUE offer two options for taking the certification exam: Candidates can take the exam online with a proctor or in person at a certified Pearson VUE test center. PSI offers online testing with a proctor but does not offer in-person testing.</p> <p>To take the exam online, certification candidates need a reliable internet connection, a webcam-enabled computer that passes a system check and a valid government-issued photo ID. For in-person testing, candidates need to schedule an appointment at a Pearson VUE test center and bring a valid photo ID.</p> <p>The SAA-C03 exam has a pass-fail designation, which means that anyone who scores at least 720 passes the exam and receive certification. AWS does not impose a maximum number of times people can take the exam, but after a failed attempt, certification candidates must wait at least 14 calendar days before scheduling a retake.</p> <p>When a candidate passes the SAA-CO3 exam and becomes certified, they receive an official <a href="https://www.techtarget.com/searchhrsoftware/feature/Guide-to-digital-badges-for-the-workplace">digital badge</a>, a printable certificate and access to certification benefits, which typically include discounts for other certifications. Amazon encourages Associate-certified cloud architects to add the badge to their LinkedIn profile, resume and email signature. The badges are linked to a easily verified digital credential issued through <a href="https://learn.credly.com/blog/heres-how-credly-helps-to-verify-your-credentials-on-the-job" target="_blank" rel="noopener">Credly</a> by Pearson.</p> <h3>Exam prerequisites</h3> <p>There are no formal prerequisites for taking the AWS Certified Solutions Architect - Associate certification exam. Anyone can register, pay and take the exam -- regardless of education or professional background. Because the exam is performance-based, however, AWS recommends certification candidates have at least one year of hands-on experience designing cloud solutions on AWS before signing up and paying the exam fee.</p> <p>Amazon also recommends that candidates who are completely new to IT or cloud computing consider earning the <a href="https://aws.amazon.com/certification/certified-cloud-practitioner/" target="_blank" rel="noopener">AWS Certified Cloud Practitioner</a> certificate first. This certification validates the certificate holder's high-level understanding of <a href="https://aws.amazon.com/products/" target="_blank" rel="noopener">AWS cloud services</a> and <a href="https://docs.aws.amazon.com/glossary/latest/reference/glos-chap.html" target="_blank" rel="noopener">terminology</a>.</p> <h3>SAA-C03 exam topics</h3> <p>The AWS Certified Solutions Architect -- Associate (SAA-C03) exam is structured around four main domains, each of which focuses on a core area of cloud architecture. Each domain is weighted to reflect its importance in cloud deployments:</p> <ol class="default-list"> <li><b>Design Secure Architectures (30%).</b> Verifies the candidate can use AWS services to design systems that protect <a href="https://www.techtarget.com/searchcio/definition/data-privacy-information-privacy">data privacy</a> and data access.</li> <li><b>Design Resilient Architectures (26%).</b> Verifies the candidate can build <a href="https://www.techtarget.com/searchdisasterrecovery/definition/fault-tolerant">fault-tolerant</a>, highly available systems using AWS tools.</li> <li><b>Design High-Performing Architectures (24%).</b> Verifies the candidate knows how to optimize performance efficiency across all system layers.</li> <li><b>Design Cost-Optimized Architectures (20%).</b> Verifies the candidate's ability to create cost-efficient systems that align with business needs.</li> </ol> <p>To pass the AWS Certified Solutions Architect - Associate exam, candidates need to have hands-on experience with a broad range of AWS tools and be familiar with a wide range of IT topics. Here are some of the most important AWS cloud services mentioned in exam questions and what they are used for.</p> <p><b>Core compute services</b></p> <ul class="default-list"> <li><b><a href="https://www.techtarget.com/searchaws/definition/Amazon-Elastic-Compute-Cloud-Amazon-EC2">Amazon Elastic Compute Cloud</a>.</b> Instance types, autoscaling, pricing models, placement groups.</li> <li><b><a href="https://www.techtarget.com/searchaws/definition/AWS-Lambda-Amazon-Web-Services-Lambda">AWS Lambda</a></b><b>.</b> Event-driven compute, function triggers, integration with other services.</li> <li><b><a href="https://www.techtarget.com/searchaws/definition/AWS-Fargate">AWS Fargate</a></b><b>.</b> Containerized application deployment and orchestration (basic understanding).</li> </ul> <p><b>Storage services</b></p> <ul class="default-list"> <li><b><a href="https://www.techtarget.com/searchaws/definition/Amazon-Simple-Storage-Service-Amazon-S3">Amazon Simple Storage Service</a> (S3).</b> Storage classes, lifecycle policies, versioning, encryption.</li> <li><b><a href="https://www.techtarget.com/searchaws/definition/Amazon-EFS-Elastic-File-System">Amazon Elastic File System</a>.</b> Use cases, performance modes, shared storage.</li> <li><b><a href="https://www.techtarget.com/searchaws/definition/Glacier-Amazon-Glacier">Amazon S3 Glacier</a></b><b>.</b> Archival storage, retrieval tiers.</li> <li><b><a href="https://www.techtarget.com/searchcloudcomputing/tip/AWS-Backup-best-practices-for-reliable-data-protection">AWS Backup</a></b><b>.</b> Centralized backup across services.</li> </ul> <p><b>Networking and content delivery</b></p> <ul class="default-list"> <li><b><a href="https://www.techtarget.com/searchcloudcomputing/definition/virtual-private-cloud-VPC">Amazon Virtual Private Cloud</a>.</b> Subnets, routing tables, Network Address Translation gateway, peering, endpoints.</li> <li><b><a href="https://www.techtarget.com/searchcloudcomputing/tutorial/How-latency-based-routing-works-in-Amazon-Route-53">Amazon Route 53</a></b><b>.</b> Domain name system routing, health checks, routing policies.</li> <li><b><a href="https://www.techtarget.com/searchcloudcomputing/answer/Cloudflare-vs-Amazon-CloudFront-Which-CDN-is-right-for-you">Amazon CloudFront</a></b><b>.</b> Content delivery network.</li> <li><b><a href="https://aws.amazon.com/elasticloadbalancing/" target="_blank" rel="noopener">Elastic Load Balancing</a></b><b>.</b> <a href="https://www.techtarget.com/searchcloudcomputing/answer/What-are-the-different-types-of-cloud-load-balancing">Cloud load balancing</a>, routing requests, health checks.</li> </ul> <p><b>Identity and security services</b></p> <ul class="default-list"> <li><b><a href="https://aws.amazon.com/iam/" target="_blank" rel="noopener">AWS Identity and Access Management</a>.</b> Policies, roles, permissions.</li> <li><b><a href="https://www.techtarget.com/searchsecurity/tip/How-to-choose-a-cloud-key-management-service">AWS Key Management Service</a>.</b> Encryption key management.</li> <li><b><a href="https://aws.amazon.com/waf/" target="_blank" rel="noopener">AWS Web Application Firewall</a>.</b> Web application security.</li> <li><b><a href="https://aws.amazon.com/cognito/" target="_blank" rel="noopener">Amazon Cognito</a>.</b> Authentication and identity federation (basic awareness).</li> </ul> <p><b>Database services</b></p> <ul class="default-list"> <li><b><a href="https://www.techtarget.com/searchcloudcomputing/tip/When-to-use-Amazon-RDS-vs-Redshift">Amazon Relational Database Service</a>.</b> Multi-Availability Zone deployments, read replicas, backup and restore.</li> <li><b><a href="https://aws.amazon.com/dynamodb/" target="_blank" rel="noopener">Amazon DynamoDB</a>.</b> NoSQL, partition keys, serverless database.</li> <li><b><a href="https://www.techtarget.com/searchcloudcomputing/answer/When-should-I-use-Amazon-RDS-vs-Aurora-Serverless">Amazon Aurora</a>.</b> High-performance relational database.</li> <li><b><a href="https://aws.amazon.com/redshift/" target="_blank" rel="noopener">Amazon Redshift</a>.</b> Data warehousing (basic awareness).</li> </ul> <p><b>Monitoring and management</b></p> <ul class="default-list"> <li><b><a href="https://aws.amazon.com/cloudwatch/" target="_blank" rel="noopener">Amazon CloudWatch</a>.</b> <a href="https://www.techtarget.com/searchnetworking/tip/Evaluate-7-cloud-monitoring-tools-for-networks">Cloud monitoring</a>, metrics, logs, alarms.</li> <li><b><a href="https://www.techtarget.com/searchsecurity/definition/AWS-CloudTrail">AWS CloudTrail</a>.</b> Application programming interface logging and auditing.</li> <li><b><a href="https://aws.amazon.com/premiumsupport/technology/trusted-advisor/" target="_blank" rel="noopener">AWS Trusted Advisor</a>.</b> Cost and security checks.</li> <li><b><a href="https://aws.amazon.com/config/" target="_blank" rel="noopener">AWS Config</a>.</b> Compliance and configuration history (basic awareness).</li> </ul> <p><b>Cost and billing</b></p> <ul class="default-list"> <li><b><a href="https://aws.amazon.com/calculator/" target="_blank" rel="noopener">AWS Pricing Calculator</a>.</b> Estimating costs.</li> <li><b><a href="https://aws.amazon.com/aws-cost-management/aws-budgets/" target="_blank" rel="noopener">AWS Budgets</a> and <a href="https://aws.amazon.com/aws-cost-management/aws-cost-explorer/" target="_blank" rel="noopener">AWS Cost Explorer</a>.</b> Tracking and <a href="https://www.techtarget.com/searchcloudcomputing/tip/Cloud-cost-management-tools-you-should-know-about">managing cloud costs</a>.</li> <li><b><a href="https://aws.amazon.com/blogs/mt/setting-up-an-amazon-cloudwatch-billing-alarm-to-proactively-monitor-estimated-charges/" target="_blank" rel="noopener">Billing Alerts</a>.</b> CloudWatch notifications for cost thresholds.</li> </ul> <p><b>Application integration and messaging</b></p> <ul class="default-list"> <li><b><a href="https://aws.amazon.com/sns/" target="_blank" rel="noopener">Amazon Simple Notification Service</a>.</b> Publish/subscribe messaging.</li> <li><b><a href="https://aws.amazon.com/sqs/" target="_blank" rel="noopener">Amazon Simple Queue Service</a>.</b> Decoupling services.</li> <li><b><a href="https://aws.amazon.com/eventbridge/" target="_blank" rel="noopener">Amazon EventBridge</a> (formerly CloudWatch Events).</b> <a href="https://www.techtarget.com/searchapparchitecture/tip/Managing-complexity-Event-driven-architecture-explained">Event-driven architecture</a>.</li> <li><b><a href="https://aws.amazon.com/step-functions/" target="_blank" rel="noopener">AWS Step Functions</a>.</b> Workflow orchestration for <a href="https://www.techtarget.com/searchitoperations/definition/serverless-computing">serverless</a> applications.</li> </ul> <p><b>Infrastructure as code</b></p> <ul class="default-list"> <li><b><a href="https://www.techtarget.com/searchaws/definition/AWS-CloudFormation-Amazon-Web-Services-CloudFormation">AWS CloudFormation</a>.</b> Template-driven <a href="https://www.techtarget.com/searchdatacenter/feature/7-must-have-cloud-infrastructure-automation-tools">cloud resource provisioning</a>.</li> <li><b><a href="https://aws.amazon.com/elasticbeanstalk/" target="_blank" rel="noopener">AWS Elastic Beanstalk</a>.</b> Simplified infrastructure as code for app deployment.</li> </ul> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/cloud_architect_skills_at_a_glance-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/cloud_architect_skills_at_a_glance-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/cloud_architect_skills_at_a_glance-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/cloud_architect_skills_at_a_glance-f.png 1280w" alt="This image provides a high-level overview of the multidisciplinary skill set required for cloud architects." height="330" width="558"> <figcaption> <i class="icon pictures" data-icon="z"></i>To be successful, cloud architects need to have broad knowledge across a wide range of IT domains. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Strategies to prepare for the AWS Certified Solutions Architect - Associate exam"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Strategies to prepare for the AWS Certified Solutions Architect - Associate exam</h2> <p>Amazon's action plan for exam preparation has four steps:</p> <ol class="default-list"> <li><b>Get to know the exam.</b> Review the <a href="https://d1.awsstatic.com/onedam/marketing-channels/website/aws/en_US/certification/approved/pdfs/docs-sa-assoc/AWS-Certified-Solutions-Architect-Associate_Exam-Guide.pdf" target="_blank" rel="noopener">official exam guide</a>, complete the certification-aligned Official Practice Question Set and take a practice test to identify knowledge gaps.</li> <li><b>Refresh AWS foundational knowledge and skills.</b> Use digital learning courses and hands-on labs to build or reinforce basic AWS skills.</li> <li><b>Review and practice for the exam.</b> Acquire a deep understanding of the exam domains by using practice questions, flashcards, instructor-led walkthroughs and simulation-style labs.</li> <li><b>Assess readiness.</b> Take a full-length official practice exam that simulates real test conditions to gauge preparedness.</li> </ol> <p>To help test-takers prepare for the exam, AWS provides two exam prep plans: one free and one paid. The free plan is available through AWS Skill Builder. It provides self-paced videos, reading materials and practice questions.</p> <p>The paid prep plan for SAA-C03 has more hours of test preparation resources. In addition to providing official practice questions that are not available in the free tier, the paid plan also provides role-based training, <a href="https://www.techtarget.com/searchhrsoftware/definition/gamification">game-based learning</a> and scenario-based experiential learning opportunities.</p> <p>In addition to using Amazon's own study resources, certification candidates can benefit from a wide range of third-party test preparation resources. For example, candidates may benefit from watching free YouTube videos about AWS or find value in taking user-created exam prep resources through a subscription on <a href="https://www.udemy.com/courses/search/?src=ukw&amp;q=AWS+Certified+Solutions+Architect+-+Associate" target="_blank" rel="noopener">Udemy</a>.</p> <p>AWS also suggests that Certified Solutions Architect - Associate certification seekers sign up for a free AWS Skill Builder account and practice taking the exam with AWS' official <a href="https://skillbuilder.aws/learn/6NV91XYP1P/official-practice-question-set-aws-certified-solutions-architect--associate-saac03--english/N1HSPV1K17" target="_blank" rel="noopener">practice exam</a>. The practice exam has 20-25 sample questions with answers and detailed explanations. It is designed to help certification candidates become familiar with the certification exam's format, question style and topics.</p> </section> <section class="section main-article-chapter" data-menu-title="AWS Certified Solutions Architect: Associate vs. Professional"> <h2 class="section-title"><i class="icon" data-icon="1"></i>AWS Certified Solutions Architect: Associate vs. Professional</h2> <p>The AWS Certified Solutions Architect - Associate certification is considered to be an intermediate certification that's one step below <a href="https://www.techtarget.com/whatis/definition/AWS-Certified-Solutions-Architect-Professional">AWS Certified Solutions Architect - Professional</a>. The Professional exam covers complex, large-scale architecture design and tests the exam-taker's ability to design enterprise-level solutions and lead <a href="https://www.techtarget.com/searcherp/tip/Learn-the-4-types-of-digital-transformation">digital transformation</a> projects that rely on AWS cloud tools.</p> <p>It's worth noting that AWS uses a recertification hierarchy. This means that, when someone receives a higher-level AWS certification, any lower-level certifications they have earned in the same track are automatically renewed. In other words, candidates who take and pass the Professional exam do not automatically receive Associate certification, but if they already hold an active Associate certification, it is automatically extended for another three years. Amazon says this policy enables professionals to maintain multiple AWS certifications without needing to retake and pay for each exam separately.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-1" src="https://www.youtube.com/embed/-xanQ3aUWms?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <p><em>Explore a list of the <a href="https://www.techtarget.com/searchcloudcomputing/tip/Top-cloud-certifications">top cloud certifications</a> for both entry-level and experienced IT professionals.</em></p> </section> AWS Certified Solutions Architect - Associate is a certification for IT professionals who use Amazon Web Services to design and implement cloud solutions. https://cdn.ttgtmedia.com/visuals/digdeeper/5.jpg https://www.techtarget.com/whatis/definition/AWS-Certified-Solutions-Architect-Associate Mon, 18 Aug 2025 09:00:00 GMT <p>Organizations that want to take advantage of machine learning capabilities require a comprehensive data preparation strategy.</p> <p>Data preparation consists of making data sets available to ML algorithms. In many cases, these algorithms need access to large amounts of data. Before these ML algorithms can access that data, it needs to be imported, processed and stored in a format suitable for analysis. This involves complex processes, as well as large storage and compute capacity.</p> <p>Here, explore some of the <a href="https://www.techtarget.com/searchcloudcomputing/answer/Compare-EMR-Redshift-and-Athena-for-data-analysis-on-AWS">key capabilities of Amazon Athena, EMR and Redshift</a> -- three data analytics services that integrate seamlessly with SageMaker AI to help IT teams navigate the data selection process. Understanding the unique strengths of each service empowers businesses to deliver more accurate, reliable ML models.</p> <section class="section main-article-chapter" data-menu-title="Select the right AWS analytics service"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Select the right AWS analytics service</h2> <p>Amazon SageMaker AI is an AWS-managed service that delivers cloud infrastructure, workflows and development tools to build, train, deploy and maintain ML models in the cloud. While SageMaker AI supports access to multiple tools for data preparation tasks, the nature of the application and its data requirements dictate the best AWS analytics service for a particular ML use case.</p> <h3>Amazon Athena</h3> <p>Athena is a query service that analyzes data files in S3 using <a href="https://www.techtarget.com/searchdatamanagement/definition/SQL">SQL</a> statements. Since it is serverless, users do not need to set up any or manage infrastructure. It is a cost-efficient option because users only pay for the queries they run. It is also a flexible service since it supports files in various formats, such as JSON, CSV, Apache ORC and Apache Parquet. It is also the best option to run ad hoc queries for data in S3.</p> <p>One common use case for Athena is log analysis to identify issues and troubleshoot. Queuing log data can also help businesses optimize their processes by analyzing performance metrics.</p> <h3>Amazon EMR</h3> <p>Amazon EMR, previously Elastic MapReduce, is a big data processing service. It launches and manages clusters that run open source data analytics frameworks, such as Apache Spark, Apache Hadoop, Apache Flink, Apache Hive and Trino. EMR can access data in a cluster's local file system, Hadoop Distributed File System (HDFS) or S3. Although EMR manages compute infrastructure using <a href="https://www.techtarget.com/searchaws/definition/Amazon-EC2-instances">EC2 instances</a>, it also supports a serverless configuration. Athena can query data using Amazon EMR, and it supports the same data formats.</p> <p>EMR provisioned clusters are a good option for jobs that require long processing tasks with a predictable workload and accessing data in HDFS or externally in S3.</p> <h3>Amazon Redshift</h3> <p>Redshift follows a data warehouse model, where extract, transform and load processes store large data sets from various sources inside a cluster. Once in the cluster, SQL statements can analyze these data sets. It is a useful tool to run queries that need to fetch and join data from multiple large tables. Redshift also manages the cluster's compute infrastructure, which is typically provisioned on EC2 instances. However, it also has the option to configure serverless compute capacity.</p> <p>Redshift is a good option for predictable, high-volume workloads with data that has been converted and stored internally in a Redshift cluster.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/key_data_preparation_steps-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/key_data_preparation_steps-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/key_data_preparation_steps-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/key_data_preparation_steps-f.png 1280w" alt="6 key steps for data preparation." height="225" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Follow these steps for successful data preparation. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Integrate AWS analytics services"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Integrate AWS analytics services</h2> <p>SageMaker Unified Studio is an integrated development environment (IDE) that gives users access to AWS' data, analytics and AI/ML capabilities in a single platform. It integrates with Athena, EMR and Redshift using its SQL extension feature to <a href="https://www.techtarget.com/searchdatamanagement/feature/Organization-and-automation-ease-data-preparation-process">ease data preparation tasks</a>. In many cases, organizations already use these services for data analytics tasks outside of SageMaker AI. This makes it easier to reuse existing infrastructure and access it for ML building and training processes.</p> <p>AWS Glue manages connections and catalogs for the data sources queried by Athena, EMR and Redshift. Users must ensure they can analyze data from their AWS analytics service using SQL statements through their IDE interfaces or SDK APIs. It's recommended to first create, execute and fine-tune these SQL statements from the AWS analytics service before running these queries from Sagemaker AI workflows.</p> <p>Remember to <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/execution-roles-and-spaces.html" target="_blank" rel="noopener">grant the required Identity and Access Management permissions</a> to the SageMaker domain that will run these data analysis tasks. These permissions must include access to relevant S3 buckets, AWS Glue catalogs and databases, as well as permissions to execute tasks in the respective AWS analytics service. Users must also configure network access, such as VPC routing and security groups, between SageMaker Unified Studio and the data analytics platform.</p> <p>The SQL extension in JupyterLab notebooks is recommended for getting started with these data analytics integrations. It provides a SQL editor UI where developers can type specific SQL commands pointing to connections and databases managed by AWS Glue. <a href="https://www.techtarget.com/searchcloudcomputing/tip/Get-started-with-Amazon-Q-Developer">Amazon Q Developer</a> is also available in JupyterLab, which is a useful generative AI-based tool that can assist and guide developers through the process.</p> <p><i>Ernesto Marquez is owner and project director at Concurrency Labs, where he helps startups launch and grow their applications on AWS. He enjoys building serverless architectures, building data analytics solutions, implementing automation and helping customers cut their AWS costs.</i></p> </section> Data preparation is crucial when building and training machine learning models with SageMaker AI. What AWS analytics services can admins use to simplify the process? https://cdn.ttgtmedia.com/rms/onlineimages/ai_a238006601.jpg https://www.techtarget.com/searchcloudcomputing/tip/Prep-data-for-machine-learning-with-AWS-analytics-services Fri, 01 Aug 2025 10:30:00 GMT <p>The artificial intelligence of things (AIoT) is the combination of <a href="https://www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence">AI</a> technologies and the internet of things (<a href="https://www.techtarget.com/iotagenda/definition/Internet-of-Things-IoT">IoT</a>) infrastructure. AIoT's goal is to create more efficient IoT operations, improve human-machine interactions, and enhance <a href="https://www.techtarget.com/searchdatamanagement/definition/data-management">data management</a> and analytics.</p> <p>AI is the simulation of human intelligence processes by machines, especially computer systems, and typically uses specialized AI algorithms, along with <a href="https://www.techtarget.com/searchenterpriseai/definition/natural-language-processing-NLP">natural language processing</a>, <a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-ML">machine learning</a> speech recognition and machine vision.</p> <p>IoT is a system of connected devices, mechanical and digital machines, or objects with unique identifiers with the ability to transfer data over a network without requiring human-to-human or human-to-computer interaction. For example, a <i>thing</i> in IoT can be a person's heart monitor implant, an automobile with built-in sensors to alert the driver when tire pressure is low, a personal assistant or any other object that can be assigned an <a href="https://www.techtarget.com/searchunifiedcommunications/definition/Internet-Protocol">Internet Protocol</a> address and transfer data over a network.</p> <section class="section main-article-chapter" data-menu-title="Why is AIoT important?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Why is AIoT important?</h2> <p>Considering the capabilities AI can add to an IoT device or supporting network, think of adding AI as raising the performance bar on IoT technology. One of the key attributes of AI is its ability to automate tasks that would otherwise be performed manually.</p> <p>A <a href="https://www.techtarget.com/iotagenda/definition/smart-home-or-building">smart home</a>, for example, has many of its functions -- e.g., security; entertainment; kitchen appliances; and heating, ventilation and air conditioning -- handled by an advanced smart device. Adding AI can help develop an all-encompassing smart home management system that, once programmed, offers performance that can literally schedule, administer and maintain all household activities, while providing an easy-to-use interface to the homeowner.</p> <p>Smart personal assistants become logical extensions of their human owners, as they enable greater capabilities and flexibility in support of their daily activities.</p> </section> <section class="section main-article-chapter" data-menu-title="How does AIoT work?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How does AIoT work?</h2> <p>In AI IoT systems and devices, AI is embedded into infrastructure components, such as programs and chipsets, which are all connected using IoT networks. Application programming interfaces (<a href="https://www.techtarget.com/searchapparchitecture/definition/application-program-interface-API">APIs</a>) are then used to ensure all hardware, software and platform components can operate and communicate without effort from the end user.</p> <p>When operational, IoT devices create and gather data, which AI analyzes to provide insights and improve efficiency and productivity. Insights are gained when an AI system allows the use of processes such as data learning.</p> <p>AIoT systems are generally designed and configured either as cloud-based or edge-based.</p> <h3>Cloud-based AIoT</h3> <p>Commonly referred to as <i>IoT cloud</i>, cloud-based IoT is the management and processing of data from IoT devices using cloud computing platforms. <a href="https://www.techtarget.com/iotagenda/tip/Know-when-to-use-cloudless-IoT-or-stick-to-the-cloud">Connecting IoT devices to the cloud</a> is essential since that's where data is stored, processed and accessed by various applications and services.</p> <p>Cloud-based AIoT is composed of the following four layers:</p> <ol class="default-list"> <li><b>Device layer.</b> This includes several types of hardware, including tags, beacons, sensors, cars, production equipment, embedded devices, and health and fitness equipment.</li> <li><b>Connectivity layer.</b> This layer comprises fields and cloud gateways consisting of a hardware or software element that links cloud storage to controllers, sensors and other intelligent devices.</li> <li><b>Cloud layer.</b> This consists of data processing via an AI engine, data storage, data visualization, data analysis and analytics, and data access via an API.</li> <li><b>User communication layer.</b> This layer is made up of web portals and mobile applications.</li> </ol> <p>Figure 1 depicts the architecture of cloud-based AIoT systems.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/cloud_based_implementation_of_aiot_architecture-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/cloud_based_implementation_of_aiot_architecture-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/cloud_based_implementation_of_aiot_architecture-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/cloud_based_implementation_of_aiot_architecture-f.png 1280w" alt="Cloud-based AIoT implementation" height="370" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Figure 1. The basic architecture of a cloud-based AIoT implementation includes these four layers. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <h3>Edge-based AIoT</h3> <p>AIoT data can also be <a href="https://www.techtarget.com/searchenterpriseai/feature/How-to-build-scalable-edge-AI-systems">processed at the edge</a>, meaning the data from IoT devices is processed as close to these devices as possible to minimize the bandwidth needed to move data, while avoiding possible delays to data analysis.</p> <p>Edge-based AIoT consists of the following three layers:</p> <ol class="default-list"> <li><b>Collection terminal layer.</b> This covers a range of hardware devices, such as embedded devices, cars, manufacturing equipment, tags, beacons, sensors, mobility devices, and health and fitness equipment, that are connected to the gateway over existing power lines.</li> <li><b>Connectivity layer.</b> This consists of the field gateways that the collection terminal layer is connected to over existing power lines.</li> <li><b>Edge layer.</b> This layer includes facilities for data storage, data processing and insight generation.</li> </ol> <p>Figure 2 depicts an edge-based AIoT implementation.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/edge_based_implementation_of_aiot_architecture-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/edge_based_implementation_of_aiot_architecture-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/edge_based_implementation_of_aiot_architecture-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/edge_based_implementation_of_aiot_architecture-f.png 1280w" alt="Edge AIoT diagram" height="286" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Figure 2. The AIoT data collected is processed closer to the source, or edge. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Applications and examples of AIoT"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Applications and examples of AIoT</h2> <p>Although many AIoT applications focus on the implementation of <a href="https://www.techtarget.com/searchenterpriseai/definition/cognitive-computing">cognitive computing</a> in consumer appliances, the following are several examples of the wider use of AIoT:</p> <figure class="main-article-image half-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/autonomous_cars-h.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/autonomous_cars-h_half_column_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/autonomous_cars-h_half_column_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/autonomous_cars-h.png 1280w" alt="How self-driving cars operate" height="274" width="279"> <figcaption> <i class="icon pictures" data-icon="z"></i>Figure 3. Autonomous cars rely on a combination of video cameras and sensor systems to collect information about adjacent vehicles, driving conditions and pedestrians. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <ul class="default-list"> <li><b>Smart cities.</b> Smart technology, such as sensors, lights and meters, collects data designed to improve operational efficiency, drive economic growth and improve residents' quality of life.</li> <li><b>Smart retail.</b> Retailers use smart cameras to recognize shoppers' faces and detect if they've scanned their items at the self-checkout before leaving the store.</li> <li><b>Smart homes.</b> Smart appliances learn through human interaction and response. AIoT appliances can also store and learn from user data to understand user habits to provide customized support.</li> <li><b>Smart office buildings.</b> AI and IoT converge in smart buildings. Companies opt for a network of smart environmental sensors installed within their offices that detect the presence of people and automatically alter the lighting and temperature to maximize energy savings. In addition, <a href="https://www.techtarget.com/searchenterpriseai/news/252516453/How-a-startup-uses-a-facial-recognition-engine-during-COVID">facial recognition technology enables smart buildings</a> to control access by using linked cameras and AI to compare live photos with a database to determine who gets access.</li> <li><b>Enterprise and industrial.</b> Manufacturers use smart chips to detect when equipment isn't functioning properly or a part needs to be replaced.</li> <li><b>Social media and human resources (HR).</b> AIoT tools can be integrated with social media and HR-related platforms to create an AI decision-as-a-service function for HR professionals.</li> <li><b>Autonomous vehicles.</b> These <a href="https://www.techtarget.com/searchenterpriseai/definition/driverless-car">vehicles</a>, as noted in Figure 3, rely on multiple video cameras and sensor systems to gather data about nearby vehicles, monitor driving conditions and look for pedestrians.</li> <li><b>Autonomous delivery robots.</b> Sensors gather data about the robot's environment -- for example, a warehouse -- and then use AI to make traversal-based decisions.</li> <li><b>Healthcare.</b> Medical devices and wearables collect and monitor real-time health data, such as heart rate, and can detect irregular heartbeats.</li> <li><b>Wearable devices.</b> <a href="https://www.techtarget.com/searchmobilecomputing/definition/wearable-technology">Wearable technology</a> can monitor and analyze personal health data to offer insights into a person's fitness, sleep and general well-being.</li> <li><b>C</b><b>ollaborative robots.</b> <a href="https://www.techtarget.com/whatis/definition/collaborative-robot-cobot">Cobots</a> are intended to assist people in the manufacturing and assembly of components. They aid humans in various tasks, such as product production, assembly, packaging and quality control, by using data from IoT devices and AI tools, including computer vision.</li> <li><b>City brains.</b> City brains are intended to promote urban development by combining machine intelligence and real-time municipal data. For example, AIoT systems can process massive logs, videos and data streams from systems and sensors throughout an urban center to detect issues such as illegal parking, road accidents and changing traffic lights.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="What are the benefits and challenges of AIoT?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What are the benefits and challenges of AIoT?</h2> <p>Benefits of AIoT include the following:</p> <ul class="default-list"> <li><b>Increased operational efficiency.</b> AI-integrated IoT devices can analyze data to reveal patterns and insights and adjust system operations to become more efficient.</li> <li><b>Ability to adjust on the fly.</b> Data can be generated and analyzed to identify points of failure, which enable the system to make adjustments as needed.</li> <li><b>Data analytics.</b> Employees don't have to spend as much time monitoring IoT devices, thus saving money.</li> <li><b>Scalability.</b> The number of devices connected to an IoT system can be increased to optimize existing processes or introduce new features.</li> <li><b>Transformational technology.</b> AIoT is transformational and mutually beneficial for both types of technology, as <a href="https://www.techtarget.com/iotagenda/tip/AI-and-IoT-How-do-the-internet-of-things-and-AI-work-together">AI adds value to IoT</a> through machine learning capabilities and improved decision-making processes. IoT adds value to AI through connectivity, signaling and data exchange. AIoT can improve businesses and their services by creating more value from IoT-generated data.</li> <li><b>Enhanced security.</b> IoT devices can be susceptible to security risks. However, AI can identify and avert these risks since AI algorithms can analyze data from sensors to discover anomalies and potential security breaches. For example, AI can analyze security camera footage to spot suspicious activity and notify security staff.</li> <li><b>Reduced human error.</b> Businesses lose millions of dollars each year as a result of <a target="_blank" href="https://www.forbes.com/sites/forbestechcouncil/2021/02/22/why-you-should-consider-the-cost-of-human-error/?sh=5c891b644a32" rel="noopener">human error</a>. By integrating machine learning with IoT technology, organizations can effectively reduce errors. In normal workflows, data must pass through multiple phases or locations, creating more opportunities for human errors, such as data entry mistakes, to occur. AIoT mitigates these risks by analyzing information at its source. Minimizing data movement and reducing the number of intermediaries involved decrease the chances of errors significantly.</li> <li><b>Personalization.</b> While IoT devices can gather information about user preferences and behavior, AI can use this information to further tailor user experiences. For example, a smart speaker can use AI to learn a user's musical preferences and generate customized playlists automatically.</li> </ul> <p>Along with its benefits and use cases, there are also instances where AIoT could fail, causing a backup in production or other negative consequences. For example, autonomous delivery robots that fail might cause a delay in the delivery of a product; smart retail stores could fail to read a customer's face, leading to the customer accidentally stealing a product; or an autonomous vehicle might fail to read its surroundings, such as an oncoming stop sign, and cause an accident.</p> <p>The following are some additional challenges associated with AIoT:</p> <ul class="default-list"> <li><b>Cybersecurity issues.</b> The growing number of devices connected through AIoT increases the risk of <a href="https://www.techtarget.com/searchsecurity/tip/6-common-types-of-cyber-attacks-and-how-to-prevent-them">cyberattacks and security breaches</a>.</li> <li><b>Complexity.</b> IoT and AI technology integration can be challenging and demand particular knowledge and abilities.</li> <li><b>Data management concerns.</b> Effective data management strategies are required for processing the data gathered from various sensors.</li> <li><b>High cost.</b> Implementing AIoT technologies can be costly due to the need for specialized equipment, software and employees.</li> <li><b>Privacy concerns.</b> There are concerns about how data acquired by AIoT devices is handled and stored, which could result in privacy issues and violations.</li> </ul> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/4FxU-xpuCww?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> </section> <section class="section main-article-chapter" data-menu-title="Standards and regulations that impact AIoT"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Standards and regulations that impact AIoT</h2> <p>Currently, the universe of activity associated with AIoT is governed by several standards, regulations and frameworks that ensure safety, privacy and ethical deployment.</p> <p>The following is a list of key initiatives.</p> <h3>AIoT regulatory initiatives</h3> <ul class="default-list"> <li><strong>E</strong><strong>uropean Union (E</strong><strong>U</strong><strong>)</strong><strong> AI Act.</strong> This 2024 legislation establishes a regulatory framework for AI systems, including AIoT.</li> <li><strong>EU Cyber Resilience Act.</strong> This 2024 legislation addresses the implementation of cybersecurity on AI-based products.</li> <li><strong>General Data Protection </strong><strong>Regulation</strong><strong>.</strong> This important 2016 EU legislation protects the collection and processing of personal data.</li> <li><strong>Network and Information Security Directive 2.</strong> This EU directive, launched in 2023, provides guidance on cybersecurity for numerous digital systems, including IoT.</li> <li><strong>California IoT</strong><strong> S</strong><strong>ecurity Law.</strong> This 2020 legislation requires device manufacturers to provide security features and unique passwords in consumer IoT devices.</li> <li><strong>U.S. IoT Cybersecurity Improvement Act.</strong> This 2020 legislation establishes cybersecurity standards for IoT devices.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="AIoT standards and frameworks"> <h2 class="section-title"><i class="icon" data-icon="1"></i>AIoT standards and frameworks</h2> <table class="main-article-table"> <thead> <tr> <td>Standard/Framework</td> <td>Purpose</td> </tr> </thead> <tbody> <tr> <td>ISO/IEC 27001 (2022)</td> <td>Fundamental standard for implementing an information security management system</td> </tr> <tr> <td>ISO/IEC 22989 (2022)</td> <td>Framework and terminology for AI system development</td> </tr> <tr> <td>ISO/IEC 23894 (2023)</td> <td>Guidance for managing risks associated with AI</td> </tr> <tr> <td>NIST SP 800-213 (2021)</td> <td>Guidance on cybersecurity for IoT devices</td> </tr> <tr> <td>NIST AI Risk Management Framework (2023)</td> <td>Framework providing guidance on how to design and develop AI systems with a focus on managing risks</td> </tr> <tr> <td>IEEE P7000 Series</td> <td>Guidance on ethical considerations in the development of advanced technology systems, such as IoT</td> </tr> <tr> <td>U.S. FCC Cyber Trust Mark (2024)</td> <td>Voluntary labeling for secure consumer IoT products</td> </tr> </tbody> </table> </section> <section class="section main-article-chapter" data-menu-title="What is the future of AIoT?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What is the future of AIoT?</h2> <p>With the integration of AI, IoT creates a much smarter system. The goal is to have these systems make accurate judgments without the need for human intervention.</p> <p><a href="https://www.techtarget.com/searchcio/definition/digital-transformation">Digital transformation</a> and the collaboration between AI and IoT have the potential to tap into unrealized customer value in several industry verticals, including edge analytics, autonomous vehicles, personalized fitness, remote healthcare, precision agriculture, smart retail, predictive maintenance and industrial automation.</p> <p>Popular and emerging trends of AIoT include the following:</p> <ul class="default-list"> <li><b>Edge computing.</b> This technology focuses on processing data in proximity to its source instead of relying on centralized cloud servers, offering benefits such as decreased <a href="https://www.techtarget.com/whatis/definition/latency">latency</a>, enhanced efficiency and reduced network congestion.</li> <li><b>Swarm intelligence.</b> Swarm intelligence involves the coordinated behavior of decentralized and self-organized systems. Inspired by natural swarms, such as bees or ants, this technology can be applied to optimize the functioning of IoT devices.</li> <li><b>5G technology.</b> One of the bigger possible innovations in AIoT is the inclusion of <a href="https://www.techtarget.com/searchnetworking/definition/5G">5G</a>. 5G is designed to enable faster transfer of large data files in IoT devices through its higher bandwidth and lower latency.</li> <li><b>Operational efficiencies.</b> AIoT could help solve existing operational problems, such as the expense associated with effective <a href="https://www.techtarget.com/searchhrsoftware/definition/human-capital-management-HCM">human capital management</a> or the complexity of supply chains and delivery models.</li> <li><b>Computer vision.</b> The goal of computer vision is to make machines comprehend and interpret visual information gleaned from the real production environment. It can analyze video streams from cameras, recognize objects and spot anomalies in AIoT applications, enabling in-the-moment automation, monitoring and optimization. Computer vision is revolutionizing the industrial sector, especially in the context of <a href="https://www.techtarget.com/searcherp/definition/Industry-40">Industry 4.0</a>, by empowering companies to improve operational efficiency, implement quality control procedures, enhance preventive maintenance practices and prioritize worker safety measures.</li> <li><strong>Smarter cities.</strong> Within urban environments, extensive use of AIoT improves vehicular traffic management, lighting, waste collection and public safety.</li> <li><strong>Enhanced simulations using digital twins. </strong>The creation of digitized replicas of physical systems helps identify potential system failures, while enhancing overall performance.</li> <li><strong>AI-enhanced cybersecurity.</strong> AI has been demonstrating how it can greatly enhance the prevention, detection and mitigation of cyberthreats.</li> <li><strong>Enhanced environmental sustainability. AIoT built into systems impacting the environment </strong><strong>is</strong><strong> expected</strong><strong> to </strong>reduce energy consumption, improve resource usage and achieve goals for the environment and sustainability.</li> </ul> <p><em>IoT can provide numerous benefits to businesses but can be challenging to deploy. Learn the prerequisites and <a href="https://www.techtarget.com/iotagenda/Ultimate-IoT-implementation-guide-for-businesses">best practices for a successful IoT installation</a>.</em></p> </section> The artificial intelligence of things (AIoT) is the combination of AI technologies and the internet of things (IoT) infrastructure. https://cdn.ttgtmedia.com/visuals/digdeeper/5.jpg https://www.techtarget.com/iotagenda/definition/Artificial-Intelligence-of-Things-AIoT Fri, 01 Aug 2025 09:00:00 GMT <p>In <a href="https://www.techtarget.com/searchoracle/definition/Oracle">Oracle</a> database management, PL/SQL is a procedural language extension to Structured Query Language (<a href="https://www.techtarget.com/searchdatamanagement/definition/SQL">SQL</a>). PL/SQL is used to combine database language and procedural programming language features and constructs so programmers can write procedural code that includes SQL easily, as though it were a single language.</p> <p>During <a href="https://www.techtarget.com/whatis/definition/compiler">compilation</a>, the Oracle Database server compiles PL/SQL program units and stores them inside the database. Since PL/SQL can include SQL statements in its syntax, PL/SQL and SQL run in the same server process during runtime to improve program execution efficiency.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineImages/oracle-oracle_database_components.jpg"> <img data-src="https://www.techtarget.com/rms/onlineImages/oracle-oracle_database_components_mobile.jpg" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineImages/oracle-oracle_database_components_mobile.jpg 960w,https://www.techtarget.com/rms/onlineImages/oracle-oracle_database_components.jpg 1280w" alt="A chart listing core components of Oracle databases" height="179" width="559"> <figcaption> <i class="icon pictures" data-icon="z"></i>PL/SQL is among the core components of Oracle databases. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p>An advantage of PL/SQL is how it accommodates writing procedural code that includes SQL so easily, as if it were a single language. PL/SQL is tightly integrated with SQL -- the most popular, widely used language for database manipulation. This integration lets programmers use all SQL statements, functions and operators. There's rarely any need to perform data type conversions with PL/SQL since the language supports all SQL data types. Also, programmers can use both <a href="https://www.techtarget.com/searchnetworking/definition/dynamic-and-static">static and dynamic</a> SQL to build more flexible, versatile applications.</p> <p>Specific application programming interface (<a href="https://www.techtarget.com/searchapparchitecture/definition/application-program-interface-API">API</a>) knowledge is not needed to map data types, prepare statements or even to process result sets. Developers can use APIs in PL/SQL to abstract complex data structures and security implementations from client applications. For example, by adding code -- inserts, updates, deletes -- to an API, they can wrap the transactional processing in that API and avoid having to create triggers, say, to implement a specific business functionality in an application. Once the API implementation is set, it can be tuned as needed. The client application remains intact during the tuning, reducing the need for application redeployment.</p> <p>With PL/SQL, business logic is built in the database through the use of blocks, with each block containing multiple SQL statements. As a result, the client code needs only a single database call per transaction. The application sends a single PL/SQL block to the database, and the database processes this entire block in one pass. This approach reduces <a href="https://www.techtarget.com/searchnetworking/tip/Assessing-performance-bottlenecks-in-virtualized-networking">network overhead</a>.</p> <p>PL/SQL allows application logic to be stored in the database itself. SQL statements and other PL/SQL constructs are grouped together and stored as a <a href="https://www.techtarget.com/searchdatamanagement/definition/schema">schema</a> object (PL/SQL procedure or function). These statements and constructs run as a unit to solve problems or perform related tasks. The centralization of application logic in the database can increase program security and developer productivity.</p> <p>Another PL/SQL benefit relates to business logic coding. In middle-tier applications, multiple interactions may occur between the application server and the database to process and execute a single business transaction. This increases overhead to the network traffic between the application and the database.</p> <p>Finally, PL/SQL is a good choice when working with Oracle Database.</p> <p>Oracle Database is a relational database management system (<a href="https://www.techtarget.com/searchdatamanagement/definition/RDBMS-relational-database-management-system">RDBMS</a>) that controls data storage, organization and retrieval in a relational database. It implements object-oriented features, like user-defined types, inheritance and polymorphism, and ensures that physical data storage is independent from logical data structures. Since it extends the relational model of a <a href="https://www.techtarget.com/searchdatamanagement/definition/database-management-system">DBMS</a> to an object-relational model, it supports the storage and manipulation of complex business models in a relational database.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/dbJMCmtLZuc?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <p>Gleaning application benefits from Oracle Database requires maintaining correct and complete data. An effective way to do this involves exposing the database using an interface that hides the implementation details, such as the tables and SQL statements. PL/SQL facilitates this. PL/SQL lets developers adopt the SmartDB paradigm, in which the code implements the surrounding business logic and the subprograms inside the database issue the SQL statements. Data can be changed and viewed only through a PL/SQL interface, ensuring correctness and completeness.</p> <section class="section main-article-chapter" data-menu-title="Blocks and procedures in PL/SQL"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Blocks and procedures in PL/SQL</h2> <p>PL/SQL is a proprietary procedural language extension to SQL. It's also a full-fledged, portable, high-performance language ideal for transaction processing.</p> <p>The basic unit of a PL/SQL source program is a block<i>.</i> Using blocks, programmers can directly write and run SQL statements in PL/SQL and add the required code to interface between the SQL statement and PL/SQL code.</p> <p>Every PL/SQL block does the following:</p> <ul class="default-list"> <li>Groups related declarations and statements with the declarations being local to the block.</li> <li>Is an executable statement,</li> <li>Is defined by the keywords DECLARE, BEGIN, EXCEPTION and END.</li> <li>Can be nested and, therefore, appear in another block wherever an executable statement is allowed.</li> <li>Can have a label.</li> </ul> <p>PL/SQL blocks can be submitted to an interactive tool and also embedded in an Oracle precompiler or <a href="https://www.techtarget.com/searchcloudcomputing/news/366552633/Oracle-databases-in-OCI-now-run-on-Microsoft-Azure">Oracle Cloud Infrastructure</a> program. Anonymous blocks are those not stored in the database and compiled every time they're loaded into memory. Block compilation always happens in three stages:</p> <ol class="default-list"> <li>Syntax checking.</li> <li>Semantic checking.</li> <li>Code generation.</li> </ol> <p>A named PL/SQL block is known as a <i>PL/SQL subprogram</i>. Subprograms, which can be either procedures or functions, can be invoked repeatedly.</p> <p>A procedure is a specific type of PL/SQL subprogram that can perform an action but not return a value. Procedures usually have parameters that pass information into subprograms when they're called and can be used in different contexts and circumstances to create more flexible subprograms.</p> <p>A PL/SQL program that is stored in a database in compiled form and can be called by name is referred to as a <a href="https://www.techtarget.com/searchoracle/definition/stored-procedure"><i>stored procedure</i></a>. Stored procedures can be used as building blocks for different applications connected to Oracle Database, and they can accept parameters at invocation.</p> <p>A PL/SQL stored procedure that is stored in the database and implicitly started when an INSERT, UPDATE or DELETE statement is issued against an associated table is called a <i>trigger</i>. PL/SQL enables programmers to specify the following:</p> <ul class="default-list"> <li>The event that occurs in the database and that the trigger will run in response to.</li> <li>When the trigger will fire -- before or after the event.</li> <li>Whether the trigger should run for each event or for each row the event affected.</li> </ul> <p>The other type of PL/SQL subprogram is known as a function. Unlike procedures, functions compute and return a value.</p> </section> <section class="section main-article-chapter" data-menu-title="How PL/SQL works"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How PL/SQL works</h2> <p>PL/SQL blocks can be variables, constants, cursors or exceptions. Blocks are defined by the keywords DECLARE, BEGIN &amp; END, and EXCEPTION, which divide the block into a declarative part, an executable part and an exception-building part, respectively.</p> <p>The declaration section of the block is written first. It is used to define and initialize constants, variables and similar items; if a variable is not initialized, it defaults to a NULL value. It is written first but is optional. The defined types, variables and similar items are manipulated in the executable part, which is the only part required to execute a PL/SQL program. The <a href="https://www.techtarget.com/searchsoftwarequality/definition/error-handling">exception-handling</a> part deals with errors that occur during execution.</p> <p>Blocks in PL/SQL, being executable statements, can be nested.</p> <p>PL/SQL is designed to compute and return a single scalar value or a single collection, such as a nested table or VARRAY. Users can create their own functions to supplement those Oracle provides. While functions can be used in an SQL statement, procedures cannot.</p> <p>The PL/SQL architecture consists of two key building blocks: the PL/SQL engine that executes procedural code and the SQL statement executor that executes SQL code. There's tight binding between SQL and PL/SQL. This is what makes PL/SQL look like a single language, enabling developers to seamlessly use all SQL statements, functions, operators and data types.</p> <p>PL/SQL program units reside in a database. When an application calls a stored procedure, Oracle Database loads the compiled program unit into the shared pool in the system global area. The PL/SQL engine, a special component of Oracle Database -- although it can also be installed in an application development tool, like Oracle Forms -- accepts any valid PL/SQL unit as input. It works with the SQL statement executor to run and process the statements in the PL/SQL procedure.</p> <p>Developers can construct several types of program units with PL/SQL: anonymous blocks, procedures, functions and packages. Anonymous blocks group related declarations and statements but are not named and are not stored in the database. Stored procedures or functions are stored in the database and can be called by name from an application. A package is a group of procedures, functions and variable definitions stored in the database and callable by other procedures, functions and variable definitions. Triggers are stored procedures associated with a database table, view or event.</p> <p>To run PL/SQL code, developers can use the following:</p> <ul class="default-list"> <li>SQL Commands page.</li> <li>Script Editor page.</li> <li>SQL Command Line.</li> </ul> <p>The SQL Commands page is the easiest option and supports reuse of SQL statements in the future since the statements can be saved as a script file in a database repository.</p> <p>PL/SQL can run with either interpreted execution or native execution. Developers can choose the method in Oracle9i and later, although native execution -- where the source code of PL/SQL program units stored in the database is compiled directly to object code -- tends to offer the best performance on computationally intensive program units.</p> </section> <section class="section main-article-chapter" data-menu-title="PL/SQL features"> <h2 class="section-title"><i class="icon" data-icon="1"></i>PL/SQL features</h2> <p>Some key features of PL/SQL include procedural elements, cursors, identifiers and embedded PL/SQL.</p> <p>PL/SQL is similar to other procedural languages in that it allows developers to define numerous types of procedural<b> </b>elements, like constants, variables and subprograms. They can also control program flow and trap and isolate <a href="https://www.theserverside.com/tip/Fix-the-5-most-common-types-of-runtime-errors-in-Java">runtime errors</a>.</p> <p>Cursors act as pointers to the context area. This is a private in-memory SQL area that stores information about processing a specific <a href="https://www.techtarget.com/searchapparchitecture/definition/parser">parsed</a> SQL statement. A PL/SQL block uses a cursor to query the database, retrieve a result set and access these records one row at a time.</p> <p>Cursors can be implicit or explicit. Oracle Database manages implicit cursors; explicit code is not required to process these cursors. Users can also use cursors explicitly using certain interfaces. Explicit cursors can be programmatically managed. They work as named resources in a PL/SQL program and enable programmers to parse embedded SQL statements effectively and better control record access in a result set.</p> <p>Implicit cursors are the default cursors in PL/SQL blocks. They're created in lieu of an explicit cursor when PL/SQL encounters a SELECT statement that returns just one row or when data manipulation language statements, like DELETE, INSERT or UPDATE, are encountered. Explicit cursors grant users better control over the context area and should be used with a SELECT statement query that returns more than one row. Both explicit and implicit cursors have the same output, even though their access is different.</p> <p>PL/SQL allows the use of identifiers to name different program items and units, including constants, variables, exceptions and subprograms. Assigning identifiers to PL/SQL blocks facilitates constructing complex applications that can be understood and maintained easily.</p> <p>PL/SQL can be embedded in high-level host languages, like <a href="https://www.techtarget.com/searchwindowsserver/definition/C">C</a>, <a href="https://www.techtarget.com/searchdatamanagement/definition/C">C++</a>, Common Business-Oriented Language (<a href="https://www.techtarget.com/searchitoperations/definition/COBOL-Common-Business-Oriented-Language">COBOL</a>) and Fortran. Oracle provides the Pro* series of precompilers (Pro*C/C++) to embed SQL and PL/SQL in application programs from these other languages. With these precompilers, developers can create highly customized applications, monitor resource use and SQL statement execution, and also use object data types in <a href="https://www.techtarget.com/searchapparchitecture/tip/Pseudocode-examples-Python-Java-JavaScript-and-C">C and C++</a> programs.</p> <p>Programming tools, like Pro*COBOL, interpret PL/SQL blocks as single, embedded SQL statements. This lets users place PL/SQL blocks anywhere in a host program where they might normally place an SQL statement. When embedding PL/SQL blocks into host programs, be sure to declare the variables that will be shared with PL/SQL and bracket the PL/SQL block with the EXEC SQL EXECUTE and END-EXEC keywords, according to Oracle documentation.</p> <p>PL/SQL includes extensive features for error handling, <a href="https://www.techtarget.com/whatis/definition/data-abstraction">data abstraction</a> and conditional compilation. When an error occurs, PL/SQL raises an exception, and normal execution automatically stops. Developers don't have to manually check every operation to confirm if it succeeded. They also don't have to get too involved with the details of data to work with its essential properties and design algorithms to manipulate the data. Finally, the conditional compilation feature means users can customize PL/SQL application functionality, say, by using new features with the latest database release or activating <a href="https://www.techtarget.com/searchsoftwarequality/definition/debugging">debugging</a> or tracing statements in the integrated development environment (<a href="https://www.techtarget.com/searchsoftwarequality/definition/integrated-development-environment">IDE</a>) without removing source text.</p> <p>After a <a href="https://www.theserverside.com/definition/Java">Java</a> stored procedure is loaded and published, users can call it. Users can connect to Oracle with the Pro*C program. Once the proper data is entered, the program automatically assigns all row values in the index-by tables to corresponding elements in the host arrays. The program repeatedly calls the procedure and displays each batch of data until no more data is found, according to Oracle's "PL/SQL User's Guide and Reference Release 2 (9.2)."</p> </section> <section class="section main-article-chapter" data-menu-title="Advantages of PL/SQL"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Advantages of PL/SQL</h2> <p>PL/SQL is built to allow programmers to use SQL in PL/SQL, meaning SQL statements can be mixed with procedural constructs. This makes it possible to use PL/SQL blocks and subprograms to group SQL statements before sending them to Oracle for execution. Users aren't required to convert between PL/SQL and SQL data types since PL/SQL fully supports SQL data types.</p> <p>The combination of SQL's data manipulation power with the processing power of procedural languages makes it easy to break down complex problems into understandable procedural code and issue SQL statements directly inside PL/SQL programs. The stored procedures in PL/SQL are callable from many different Oracle Database clients, including Pro*C or Oracle Call Interface, and from Oracle Reports and Oracle Forms. The procedural code can also be reused across multiple applications, and PL/SQL variables can be interchanged with data inside a table.</p> <p>Without PL/SQL, Oracle must process SQL statements one at a time. In a network environment, this can affect traffic flow and slow down response time. PL/SQL blocks can be compiled once, stored in executable form, run in the database server, and invoked repeatedly to <a href="https://www.techtarget.com/searchsoftwarequality/tip/Acceptable-application-response-times-vs-industry-standard">improve response time</a> and lower memory requirements and invocation overhead. Developers can also issue a SQL query and process the result set one row at a time.</p> <p>PL/SQL enables users to send a block of statements to the database, reducing traffic significantly between the app and database. Performance is also improved because Oracle Database reuses the SQL statements every time the same code runs.</p> <p>Since PL/SQL is a portable language, PL/SQL programs can run on any operating system or platform where Oracle Database runs.</p> <p>When stored in subprograms, PL/SQL increases <a href="https://www.techtarget.com/searchdatacenter/definition/scalability">scalability</a> by centralizing access processing on the database server. The shared server's memory facilities enable Oracle Database to support multiple concurrent users on a single node. Scalability can be improved further by <a href="https://www.techtarget.com/searchnetworking/definition/multiplexing">multiplexing</a> network connections using Oracle Connection Manager -- a router that provides features like session multiplexing, access control and protocol conversion useful for streamlining the sending of client connection requests, either to the next hop or directly to the database server.</p> <p>Users need to maintain only one copy of a subprogram on the database server instead of multiple copies on each client system. This makes it easier to manage the programs and even alter subprograms without affecting the apps that use or invoke them.</p> <p>PL/SQL also enables users to create applications that generate webpages directly from the database. This enables users to make their databases available on the web.</p> <p>PL/SQL Server Pages enable users to develop webpages and are an alternative to coding a stored subprogram that writes the <a href="https://www.theserverside.com/definition/HTML-Hypertext-Markup-Language">Hypertext Markup Language</a> code one line at a time. During development, PL/SQL Server Pages are used as templates, and users can design layouts and write PL/SQL scripts to generate the content.</p> </section> <section class="section main-article-chapter" data-menu-title="PL/SQL vs. SQL"> <h2 class="section-title"><i class="icon" data-icon="1"></i>PL/SQL vs. SQL</h2> <p>Although they have similar-sounding names and PL/SQL is a procedural extension of SQL, SQL and PL/SQL are significantly different programming languages.</p> <p>Unlike SQL, which is an industry-standard database language, PL/SQL is unique to Oracle and is meant to be used with Oracle Database. This means that the Oracle Database server compiles PL/SQL program units and the units themselves are stored inside the database. PL/SQL includes all the features of programming languages and integrates easily with SQL.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/oracle-plsql_vs_sql.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/oracle-plsql_vs_sql_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/oracle-plsql_vs_sql_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/oracle-plsql_vs_sql.png 1280w" alt="A chart listing 6 important distinctions between PL/SQL and SQL" height="427" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Six important distinctions between PL/SQL and SQL </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p>PL/SQL includes support for <a href="https://www.techtarget.com/searchapparchitecture/definition/object-oriented-programming-OOP">object-oriented programming</a>. PL/SQL supports this programming paradigm using abstract data types. ADTs separate large systems into logical components to reduce programming complexity. Each ADT consists of a data structure, which itself is composed of variables, known as <i>attributes</i>, and subprograms that manipulate the data. In PL/SQL, ADTs are known as <i>user-defined types</i> and <i>object types</i> and are stored in the database.</p> <p>SQL does not support variables or control structures, like FOR loops. PL/SQL supports both. Using control structures, users can manipulate and process Oracle data using conditional, iterative and sequential flow-of-control statements, like IF-THEN and FOR loops.</p> <p>PL/SQL is primarily an application-oriented language suitable for applications like database administration automation, <a href="https://www.techtarget.com/searchdatacenter/tip/XML-vs-YAML-Compare-configuration-file-formats">Extensible Markup Language</a> document management and integration of databases with webpages. SQL is a data-oriented language whose chief purposes are to help users store, retrieve, manipulate and manage data in relational databases and to create and modify database structures.</p> <p>How queries are processed in SQL vs. PL/SQL is another distinguishing difference. In SQL, the DBMS processes a query as single statement via a series of steps to retrieve data. In PL/SQL, blocks of statements are sent to the database with only a single database call from the client code per transaction, reducing network traffic.</p> <p>SQL and PL/SQL use different syntax. Where SQL consists of only pure SQL statements, PL/SQL combines SQL statements with procedural logic, so programmers can write procedural code that includes SQL as if it were a single language.</p> <p>Other differences between SQL and PL/SQL include the following:</p> <ul class="default-list"> <li>SQL does not provide error and exception handling, while PL/SQL does.</li> <li>SQL can be embedded into PL/SQL, while the reverse is not possible.</li> <li>SQL interacts directly with the database server, while PL/SQL does not.</li> <li>PL/SQL offers high processing speeds even for high-volume data; speeds are limited for high-volume data in SQL.</li> </ul> <p>While SQL is considered a source of data for reports, webpages and screens, PL/SQL is viewed as an application language similar to Java and <a href="https://www.techtarget.com/searchapparchitecture/tip/PHP-8-features-that-prove-its-for-more-than-just-web">PHP</a>. PL/SQL can be used to build, format and display reports, webpages and screens.</p> </section> <section class="section main-article-chapter" data-menu-title="PL/SQL Developer IDE"> <h2 class="section-title"><i class="icon" data-icon="1"></i>PL/SQL Developer IDE</h2> <p>Regardless of which programming language is used to write programs, programmers need a development environment and tools to help them execute and streamline many tasks. These tools may include a coder, compiler, interpreter and debugger. Such an environment is called an <i>integrated development environment</i>.</p> <p>One popular IDE for PL/SQL is called PL/SQL Developer. Ideal for developing software -- stored program units -- for the Oracle Database environment, PL/SQL Developer is developed and built by Netherlands-based company Allround Automations. This IDE provides a powerful PL/SQL editor with useful features, like syntax highlighting, compiler hints, refactoring, code contents, bookmarks, multilevel undo and redo, and hyperlink navigation, all of which help to streamline and simplify PL/SQL coding, even for demanding tasks.</p> <p>A debugger tool is integrated with PL/SQL Developer, offering features like view call stack, run until exception, and view and set variables. Programmers can debug program units without modifying the debugger. Easy to use, the debugger can help fix a wide range of bugs. It also enables programmers to find the cause of exceptions by directing them to the line that caused the exception and displaying the variable values at the time of the exception.</p> <p>PL/SQL Developer includes AI Assistant, which provides numerous predefined AI Functions to help programmers create, modify, finish, fix and optimize PL/SQL code quickly. Developers can also modify the predefined AI Functions, ask general questions to AI Assistant about code and create new AI Functions for specific tasks.</p> <p>Another unique feature of PL/SQL Developer is PL/SQL Beautifier. Developers can work with a user-defined set of rules to format their SQL and PL/SQL code. As soon as they compile, save or open a file, PL/SQL Developer automatically beautifies it, thus improving its readability and understandability. This is crucial to keep the <a href="https://www.techtarget.com/searchsoftwarequality/definition/software-development-life-cycle-SDLC">software development lifecycle</a> moving and to facilitate easy collaboration among members of large project teams.</p> <p>Since it is a multithreaded IDE, PL/SQL Developer lets developers continue working even when other tasks run, such as SQL queries, PL/SQL programs or debug sessions. There's no need to install a database object to use PL/SQL Developer, and the only <a href="https://www.techtarget.com/searchapparchitecture/definition/middleware">middleware</a> needed is Oracle Net.</p> <p><em>There are several different database options for organizations to choose from. Learn more about how <a href="https://www.techtarget.com/searchdatamanagement/answer/What-is-the-difference-between-DBMS-and-RDBMS">DBMS and RDBMS compare</a>. Also, NoSQL systems are becoming more prevalent in the cloud, with numerous cloud providers and vendors offering them. Learn about the various <a href="https://www.techtarget.com/searchcloudcomputing/tip/Compare-NoSQL-database-types-in-the-cloud">types of NoSQL databases</a> and their pros and cons.</em></p> </section> In Oracle database management, PL/SQL is a procedural language extension to Structured Query Language (SQL). https://cdn.ttgtmedia.com/visuals/digdeeper/5.jpg https://www.techtarget.com/searchoracle/definition/PL/SQL Wed, 30 Jul 2025 09:00:00 GMT <p>Cloud services create many new challenges for system and network administration, as well as cybersecurity. Managing and securing distributed data and processes is far more complex than managing on-premises network environments with strict boundaries and complete control of resources.</p> <p>These new challenges led to the cloud shared responsibility model. Under this framework, service providers and customers maintain two separate facets of cloud management and security. While cloud service providers (CSPs) provide the services and security <i>of</i> the cloud, customers retain management <i>within</i> the cloud. The model differentiates the responsibilities for specific aspects of cloud administration.</p> <p>Let's explore the shared responsibility model, including its advantages and disadvantages. We'll also review examples of responsibilities, documents and different approaches by the major CSPs.</p> <section class="section main-article-chapter" data-menu-title="The shared responsibility model"> <h2 class="section-title"><i class="icon" data-icon="1"></i>The shared responsibility model</h2> <p>The <a href="https://www.techtarget.com/searchcloudcomputing/definition/shared-responsibility-model">cloud shared responsibility model</a> -- sometimes called the shared security model -- divides responsibilities between the customer and CSP. This clear delineation covers all aspects of configuration, security and maintenance, reducing risks and improving an organization's overall availability and security.</p> <p>The shared responsibility model emerged from the growth of the cloud computing industry and became standard practice by about 2010. AWS was one of the first major cloud service providers to <a target="_blank" href="https://aws.amazon.com/compliance/shared-responsibility-model/" rel="noopener">clearly define the responsibilities</a>, especially with the growth of IaaS.</p> <p>Let's start by looking at the shared responsibilities between CSPs and their customers. Customer responsibilities include the following:</p> <ul class="default-list"> <li>Identity and access management.</li> <li>Data protection, which includes backups, encryption and access controls</li> <li>Application security and patching.</li> <li>Cloud resource configuration for VMs, containers, webapps and virtual networks.</li> <li>Endpoint and network access security.</li> </ul> <p>The customer is responsible for access to the data in the cloud, which pertains to user account management, permissions, backups, resource configuration and application management. It also encompasses the user systems and networks that access data in the cloud.</p> <p>Cloud provider responsibilities include the following:</p> <ul class="default-list"> <li>Physical security and availability of data centers.</li> <li>Hardware, network and infrastructure supporting cloud services, including patching.</li> <li>Virtualization layers.</li> <li>Data centers and infrastructure where customer resources reside.</li> </ul> <p>Some organizations outsource cloud computing support and administration to third-party providers or even the primary CSPs. In those cases, the providers manage some or all customer responsibilities.</p> <p>Responsibilities also vary by service type, with cloud providers taking on more management tasks in SaaS deployments while customers do more with IaaS services.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/cloud_computing-cc_101_shared_responsibility_model-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/cloud_computing-cc_101_shared_responsibility_model-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/cloud_computing-cc_101_shared_responsibility_model-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/cloud_computing-cc_101_shared_responsibility_model-f.png 1280w" alt="Breakdown of the cloud shared responsibility model." height="364" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Use this guide to understand consumer and provider responsibilities under the shared responsibility model for the three main cloud service models. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="CSP responsibilities"> <h2 class="section-title"><i class="icon" data-icon="1"></i>CSP responsibilities</h2> <p>The specific responsibilities of the CSP vary depending on the service model. Here are the general aspects CSPs manage:</p> <h3>SaaS</h3> <p>For SaaS, the CSP retains responsibility for the entire supporting application infrastructure, including hardware, OS and the application itself. Using Windows 365, Microsoft maintains the data centers and systems. Microsoft also upgrades and patches the applications that make up the service.</p> <h3>PaaS</h3> <p>PaaS models are similar, though application configuration and management are customer responsibilities. The CSP retains control of the infrastructure and OS.</p> <h3>IaaS</h3> <p>IaaS deployments consist of the CSP managing the hardware and virtualization layers, including drivers, device failures, device capabilities and virtualization software or container engines. These functions remain hidden from the cloud customer.</p> </section> <section class="section main-article-chapter" data-menu-title="Cloud customer responsibilities"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Cloud customer responsibilities</h2> <p>Organizational responsibilities also vary by service model.</p> <p>Generally, anything not covered by the CSP's responsibilities falls onto individual cloud management staff, including:</p> <h3>SaaS</h3> <p>Customers relying on SaaS services must carefully <a href="https://www.techtarget.com/searchcloudcomputing/definition/subscription-based-pricing-model">monitor subscriptions</a> and application access. These customers can also develop specific application customizations. Continuing the example of Windows 365 from above, customers maintain control over which employees access services.</p> <h3>PaaS</h3> <p>Developers, database administrators and other users who need PaaS access retain control over the data generated by the platform. The business also retains access control, ensuring only authorized employees can use the PaaS service.</p> <h3>IaaS</h3> <p>A business's IT staff controls the VMs making up the IaaS platform. Administrators define VMs, install OSes, build networking solutions and create any custom configurations. Note that this team is also responsible for OS and application patching. The business also retains control of data production at this level.</p> <p>Keep in mind that IT staff might need to develop <a href="https://www.techtarget.com/whatis/feature/Top-20-cloud-computing-skills-to-boost-your-career">cloud-specific skills</a> to handle the responsibilities because they differ from on-premises deployments. Hybrid and multi-cloud deployments vary even further. Engaging third-party and cloud service providers' technical services helps with these tasks.</p> <p><i>Damon Garn owns Cogspinner Coaction and provides freelance IT writing and editing services. He has written multiple CompTIA study guides, including the Linux+, Cloud Essentials+ and Server+ guides, and contributes extensively to Informa TechTarget, The New Stack and CompTIA Blogs.</i></p> </section> The shared responsibility model is an integral part of managing any cloud infrastructure. Review your cloud deployment today to ensure your business fulfills its responsibilities. https://cdn.ttgtmedia.com/rms/onlineimages/check_g1182594295.jpg https://www.techtarget.com/searchcloudcomputing/tip/What-the-cloud-shared-responsibility-model-requires-of-you Fri, 25 Jul 2025 12:19:00 GMT <p>A distributed database is a database that consists of two or more files located in different sites on the same or different networks. Processing is distributed among multiple <a href="https://www.techtarget.com/searchdatamanagement/definition/database">database</a> nodes stored in multiple physical locations.</p> <p>In effect, the data in a distributed database is stored across multiple computers (physical or <a href="https://www.techtarget.com/searchitoperations/definition/virtual-machine-VM">virtual machines</a>) so a version of the database runs on multiple locations. Each of these locations, which can be geographically separated, is known as a <i>node</i> or <i>instance</i>. To users, however, the database looks like one single database.</p> <p>Distributed databases are <a href="https://www.techtarget.com/searchdatacenter/definition/scalability">scalable</a>, highly available, deliver high performance, and can accommodate different data types. <a href="https://www.techtarget.com/searchdisasterrecovery/definition/data-replication">Data replication</a> and fragmentation are two ways in which data is stored in multiple sites to create a distributed database and ensure its high availability.</p> <section class="section main-article-chapter" data-menu-title="Uses of distributed databases"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Uses of distributed databases</h2> <p>Distributed databases are suitable for scenarios and applications that require the following:</p> <ul class="default-list"> <li>High performance.</li> <li>Data replication and redundancy.</li> <li><a href="https://www.techtarget.com/searchnetworking/definition/load-balancing">Load balancing</a>.</li> <li>Resilience.</li> <li>Fault tolerance.</li> <li><a href="https://www.techtarget.com/searchdatabackup/definition/backup">Data backup</a>.</li> <li>Continuous database monitoring and failure detection.</li> </ul> <p>Distributed databases are ideal for applications that require high <a href="https://www.techtarget.com/searchstorage/definition/data-availability">data availability</a> and <a href="https://www.techtarget.com/searchdisasterrecovery/definition/fault-tolerant">fault tolerance</a>, such as <a href="https://www.techtarget.com/searchcio/definition/e-commerce">e-commerce</a> and corporate management information systems, financial transactions, online games, multimedia applications and more.</p> <p>Distributed databases are particularly useful when users must access data remotely or from a variety of mobile devices. Applications that use large or diverse data sets can also benefit from distributed databases. These might include healthcare, military, and internet of things (<a href="https://www.techtarget.com/iotagenda/definition/Internet-of-Things-IoT">IoT</a>) applications.</p> </section> <section class="section main-article-chapter" data-menu-title="How does a distributed database work?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How does a distributed database work?</h2> <p>In a distributed database, data is placed on multiple servers or computers consisting of individual nodes. The nodes might be physical systems or virtual machines, and they might be geographically separated from each other without sharing physical components. Each of these nodes stores a copy of the data set and runs on database management system (<a href="https://www.techtarget.com/searchdatamanagement/feature/Evaluating-the-different-types-of-DBMS-products">DBMS</a>) software that provides centralized control and <a href="https://www.computerweekly.com/opinion/Too-many-data-sources-and-too-little-consistency">consistency</a> across the distributed environment and manages tasks like data partitioning, replication, and updates. To ensure consistency between the replicas or copies of the data between multiple nodes, those nodes use an <a href="https://www.techtarget.com/whatis/definition/algorithm">algorithm</a>, such as the <i>Raft consensus algorithm</i>, to achieve consensus. This means that the replicas agree beforehand that the data being entered is correct before a write is committed.</p> <p>Once the nodes are set up, they can receive read/write requests. Due to the <a href="https://www.techtarget.com/whatis/definition/consensus-algorithm">consensus mechanism</a>, they avoid consistency problems. The nodes communicate with each other using one of three communication methods:</p> <ul class="default-list"> <li><b>Unicast.</b> This is when one node sends a message only to one other node.</li> <li><b>Broadcast.</b> This happens when a node sends a message to all other nodes in the distributed database.</li> <li><b>Multicast.</b> Here, a message is sent only to some nodes in the distributed database.</li> </ul> <p>A distributed database improves data availability and sharing by spreading or "replicating" data across multiple locations. It supports distributed transactions with <a href="https://www.techtarget.com/searchdatamanagement/definition/ACID">ACID</a> (atomicity, consistency, isolation, durability) properties involving more than one node and also allows for horizontal scaling.</p> <p>Data partitioning is essential for proper data distribution and optimal user access to that data. The database can either partition data <i>horizontally</i> or <i>vertically</i>. In horizontal data partitioning, the data tables are split into rows across multiple nodes. In contrast, vertical data partitioning involves splitting the tables into columns across all the nodes in the database. Either way, the resulting data sets are known as <a href="https://www.techtarget.com/searchoracle/definition/sharding"><i>shards</i></a>.</p> </section> <section class="section main-article-chapter" data-menu-title="Features of distributed databases"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Features of distributed databases</h2> <p>When in a collection, distributed databases are logically interrelated and often represent a single logical database. Data is physically stored across multiple sites and managed independently by the various nodes. This approach -- placing the same data on multiple servers or computers -- is called replication<i>.</i> This feature increases data availability and minimizes the potential for data loss.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/data_management-database_replication.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/data_management-database_replication_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/data_management-database_replication_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/data_management-database_replication.png 1280w" alt="A diagram illustrating how database replication works." height="304" width="520"> <figcaption> <i class="icon pictures" data-icon="z"></i>Database replication ensures the data in distributed databases remains up to date. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p>Distributed databases also include these features:</p> <ul class="default-list"> <li><b>Distributed query processing.</b> The presence of multiple nodes and replicated data supports distributed and faster query processing.</li> <li><b>Distributed transaction management. </b>A single transaction can involve more than one node, with each site managing its transactions with other sites.</li> <li><b>Hardware- and OS-independent</b>. Heterogeneous distributed databases use different machines or sites that use different hardware and operating systems (<a href="https://www.techtarget.com/whatis/definition/operating-system-OS">OSes</a>). They might also contain different data <a href="https://www.techtarget.com/searchdatamanagement/definition/schema">schemas</a>. While these differences can increase database complexity, they offer greater flexibility in the types of data that can be stored while yielding better performance and scalability.</li> <li><b>Support for ACID transactions.</b> In distributed databases, all transactions are treated as single units (atomicity); data consistency is maintained by enforcing predefined rules and data constraints (consistency); transactions are isolated from each other to prevent data conflicts and maintain data integrity (isolation); and data is preserved even if a system fails (durability).</li> <li><b>Multiple communication means. </b>Nodes can communicate with each other using unicast, broadcast or multicast communication.</li> <li><b>High fault tolerance. </b>Distributed databases use numerous processes to increase tolerance to failures or operational interruptions. These processes include data replication, data backup, <a href="https://www.techtarget.com/searchdisasterrecovery/tip/High-availability-and-resiliency-A-DR-strategy-needs-both">continuous failure detection</a> and load balancing.</li> <li><b>Query optimization.</b> Distributed databases use techniques like cost-based query optimization to efficiently distribute queries across nodes, execute queries and minimize data transfer traffic between nodes.</li> <li><b>Failure detection.</b> Distributed database systems detect issues like failed nodes and data tampering through continuous monitoring techniques, such as data and watchdog timers. These issues might result from technical problems, <a href="https://www.techtarget.com/searchsecurity/tip/6-common-types-of-cyber-attacks-and-how-to-prevent-them">cyberattacks</a> or natural disasters.</li> </ul> <p>A common misconception is that a distributed database is a loosely connected file system. It's actually much more complicated. Distributed databases incorporate multiple nodes to replicate data and software to facilitate communication between these nodes. They also incorporate transaction processing (ACID-compliant) but are not synonymous with transaction processing systems.</p> </section> <section class="section main-article-chapter" data-menu-title="Distributed database vs. centralized database"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Distributed database vs. centralized database</h2> <p>A centralized database consists of a single database file located at one site using a single network. The database is stored, modified and managed from a single location, and the location is accessed via an internet connection. It's easy to access and coordinate data in a centralized database since all the data is stored at one location. Data redundancy is also minimal, and costs are quite low, since there's no need to manage multiple locations or data sets.</p> <p>A centralized database requires a lot of management effort. Data traffic also tends to be high, which can reduce database performance and create <a href="https://www.techtarget.com/whatis/definition/latency">latency</a> issues for users. Another important drawback of centralized databases is that they are vulnerable to <a href="https://www.techtarget.com/searchdatacenter/definition/Single-point-of-failure-SPOF">single points of failure</a>. In case of a failure event, such as a cyberattack, the entire database goes down, and the data within it can also be destroyed permanently.</p> <p>By contrast, a distributed DBMS consists of multiple nodes that store data and are spread across multiple physical locations. The nodes communicate with each other using the network to facilitate data querying and access. The database integrates data logically so it can be managed as if it were all stored in the same location. Also, the DBMS periodically <a href="https://www.techtarget.com/searchdisasterrecovery/definition/synchronous-replication">synchronizes all the data</a> and ensures that data updates and deletes performed at one location are automatically reflected in the data stored elsewhere.</p> <p>Distributed databases are more flexible and scalable than centralized databases. The distribution of data across multiple nodes, plus built-in fault tolerance and continuous database monitoring, increases data availability, reduces latency, and improves database performance. The risk of a single point of failure is also much lower, providing greater resiliency and ensuring continuous operations.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/whatis-databases.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/whatis-databases_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/whatis-databases_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/whatis-databases.png 1280w" alt="A chart covering the primary differences between distributed databases and centralized databases" height="496" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Five ways centralized databases differ from distributed databases. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Types of distributed databases"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Types of distributed databases</h2> <p>There are two main types of distributed databases: <i>homogeneous</i> and <i>heterogeneous</i>.</p> <h3>Homogeneous distributed database</h3> <p>A homogeneous distributed database encompasses different sites (nodes/machines) that all store the same data. They also have the following characteristics:</p> <ul class="default-list"> <li>Use the same <a href="https://www.techtarget.com/searchdatamanagement/definition/data-modeling">data model</a>.</li> <li>Work with the same OS.</li> <li>Use the same DBMS across all sites.</li> </ul> <p>A homogeneous distributed database is easier to manage because the nodes in it are all similar and store data in an identical manner. It also offers significant protection from data loss due to built-in redundancy.</p> <p>Other benefits include scalability, high availability and improved <a href="https://www.techtarget.com/searchcloudcomputing/definition/data-residency">data residency</a> control. They can offer high performance even when handling large data sets and high-volume transactions.</p> <p>Homogeneous distributed database systems appear to the user as a single system, and they can be much easier to design and manage. For a distributed database system to be homogeneous, the data structures and database application at each location must be either identical or compatible. Homogeneous databases can be <i>autonomous</i> or <i>nonautonomous</i>.</p> <p>An autonomous distributed database<i> </i>consists of numerous database instances or nodes that are physically separate and work independently with their own complete set of data. These nodes only need a single centralized managed layer for centralized provisioning and universal updates (across all nodes) and to provide a unified logical view to applications. Autonomous databases provide particularly high availability, improved performance, and enhanced scalability for large data sets and high-volume transactions.</p> <p>Unlike autonomous distributed database<i>s, </i>nonautonomous homogeneous distributed databases<i> </i>rely on centralized control using a single DBMS. While data is distributed across multiple nodes, the DBMS coordinates data partitioning, distribution, storage and retrieval. The DBMS also handles various complexities related to updates, communications, replication and ensures consistency across all nodes.</p> <h3>Heterogeneous distributed database</h3> <p>The other main distributed database type is a<b> </b>heterogeneous database. In<b> </b>these databases, multiple sites or machines house different data sets. They also use different schemas, Oses and database applications. The sites might not even be aware of each other. Different nodes can have different hardware, software and data structure, or might be in locations that are not compatible. Users at one location might be able to read data at another location but not upload or alter it. These differences can create problems during query processing and transactions can require translations between sites. Heterogeneous distributed databases also can be difficult to use, with associated costs prohibitive for many businesses.</p> <p>Despite these complexities, heterogeneous distributed databases offer several benefits, including greater flexibility in data models and schema choices. Multiple nodes can function independently but also work together to respond appropriately to and process a query. This <i>data virtualization</i> is a key feature of heterogenous distributed databases of the <i>federated</i> type. An <i>unfederated</i> heterogeneous distributed database also consists of nodes that operate independently. But unlike federated databases, unfederated databases rely on a centralized application to coordinate data distribution, access, control, and updates.</p> </section> <section class="section main-article-chapter" data-menu-title="Types of data in distributed databases"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Types of data in distributed databases</h2> <p><a href="https://www.techtarget.com/searchdisasterrecovery/definition/data-replication">Replicated data</a> is used to create instances of data in different parts of the database. By using replicated data, distributed databases can access identical data locally, avoiding traffic. Replicated data can be divided into two categories: read-only and writable data.</p> <p>Read-only versions of replicated data allow revisions only to the first instance; subsequent enterprise data replications are then adjusted. Writable data can be altered, but the first instance is changed immediately.</p> <p>Horizontally fragmented data involves the use of primary keys that refer to one record in the database. Horizontal fragmentation is usually reserved for situations in which business locations only need to access the database pertaining to their specific branch.</p> <p>Vertically fragmented data involves using copies of <a href="https://www.techtarget.com/searchdatamanagement/definition/primary-key">primary keys</a> available in each section of the database and accessible to each branch. Vertically fragmented data is utilized when the branch of a business and the central location interact with the same accounts in different ways.</p> <p>Reorganized data is data that has been adjusted or altered for decision support databases. Reorganized data is typically used when two different systems are handling transactions and decision support. <a href="https://www.techtarget.com/searchcio/definition/decision-support-system">Decision support systems</a> can be maintenance-intensive and online transaction processing might require reconfiguration when many requests are being made.</p> <p>Separate schema data partitions the database and the software used to access it to fit different departments and situations. There is usually an overlap between different databases in separate schema data.</p> </section> <section class="section main-article-chapter" data-menu-title="Advantages of distributed databases"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Advantages of distributed databases</h2> <p>There are many advantages to using distributed databases.</p> <p>Data replication across multiple nodes increases data availability and minimizes database failures.</p> <p>Distributed databases accommodate modular development, meaning systems can be expanded by adding new computers and local data to the new site and connecting them to the distributed system without interruption.</p> <p>Also, different data structures and schemas can be used. Such scalability and flexibility are critical for modern applications that rely on vast data volumes and different data types that change at high velocities.</p> <p>Distributed databases incorporate multiple mechanisms to enable proactive and continuous failure monitoring and investigations. In addition, there is no single point of failure since the data is placed in multiple nodes, increasing <a href="https://www.techtarget.com/searchdisasterrecovery/tip/High-availability-and-resiliency-A-DR-strategy-needs-both">resiliency</a>. Even if a component fails, the distributed system will continue to function, albeit at reduced performance, until the error is fixed. In contrast, a single failure in a centralized database can halt the entire system.</p> <p>Built-in load balancing improves database performance, minimizes system inefficiency, and reduces user wait times. Distributed databases also offer query optimization<b> </b>to speed up query processing and reduce data transfer traffic between nodes.</p> </section> <section class="section main-article-chapter" data-menu-title="Disadvantages of distributed databases"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Disadvantages of distributed databases</h2> <p>An important disadvantage of a distributed database is cost. Multiple servers or computer clusters are needed to replicate and distribute data, creating a need for additional hardware and networking elements. This increases complexity and cost.</p> <p>Data consistency is another concern. Maintaining data consistency requires additional effort with multiple sites and data schemas.</p> <p>Latency can be an issue when users query the database from multiple nodes. It's important to consider these issues during database design and to manage the database carefully to ensure <a href="https://www.techtarget.com/searchdatabackup/definition/restore">data restoration</a> (in case of failure) and maintain data security (to prevent data loss and data integrity degradation).</p> </section> <section class="section main-article-chapter" data-menu-title="Examples of distributed databases"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Examples of distributed databases</h2> <p>Though there are many distributed databases to choose from, some prominent examples include the following:</p> <ul class="default-list"> <li><a href="https://www.techtarget.com/whatis/definition/Apache">Apache</a> Ignite.</li> <li><a href="https://www.techtarget.com/searchdatamanagement/news/252528398/Apache-Cassandra-41-extends-open-source-NoSQL-database">Apache Cassandra</a>.</li> <li>Apache HBase.</li> <li><a href="https://www.techtarget.com/searchdatamanagement/news/366619602/Couchbase-integrates-with-Nvidia-NIM-to-aid-AI-development">Couchbase Server</a>.</li> <li><a href="https://www.techtarget.com/whatis/definition/Amazon">Amazon</a> SimpleDB.</li> <li>FoundationDB.</li> </ul> <p>Apache Ignite is a distributed database for high-performance applications. It offers multi-tier storage and supports distributed ACID transactions. Ignite can be used as a traditional Structured Query Language (<a href="https://www.techtarget.com/searchdatamanagement/definition/SQL">SQL</a>) database by leveraging Java Database Connectivity (<a href="https://www.theserverside.com/definition/Java-Database-Connectivity-JDBC">JDBC</a>) drivers or Open Database Connectivity (<a href="https://www.techtarget.com/searchoracle/definition/Open-Database-Connectivity">ODBC</a>) drivers. Users can also use native SQL APIs for numerous programming languages (<a href="https://www.theserverside.com/definition/Java">Java</a>, <a href="https://www.techtarget.com/whatis/definition/Python">Python</a>, <a href="https://www.techtarget.com/whatis/definition/C-Sharp">C#</a> and more) to execute custom tasks across the database. Additionally, developers can deploy continuous queries in any of these languages and process streams of changes on both the database and application side. Ignite integrates with TensorFlow to allow users to build scalable <a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-ML">machine learning</a> models.</p> <p>Apache Cassandra is an open source <a href="https://www.techtarget.com/searchdatamanagement/definition/NoSQL-Not-Only-SQL">NoSQL</a> distributed database that uses multiple identical nodes, making it ideal for mission-critical data and applications that cannot afford to lose data.<b> </b>It features a masterless and elastic architecture to minimize the possibility of data loss in case of a data center outage. In addition, data replication across multiple data centers eliminates single points of failure and ensures high fault tolerance and low latency. Even if nodes do fail, they can be easily replaced to minimize downtime. Cassandra offers both synchronous and asynchronous replication for updates, provides an audit logging feature for enhanced security and observability, and is highly scalable. It also provides a tool to capture and replay production workloads for analysis.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/Vd6jCNa5E24?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <p>Apache HBase is a scalable, distributed <a href="https://www.techtarget.com/searchdatamanagement/definition/Hadoop">Hadoop</a> database modeled after Google's Bigtable (a distributed storage system for structured data). HBase runs on top of the HDFS (Hadoop Distributed File System) and supports structured data storage for very large tables that can contain billions of rows and millions of columns. This versioned, nonrelational database is ideal for applications that need random, real-time access (read/write) to <a href="https://www.techtarget.com/searchdatamanagement/definition/big-data">big data</a>. It offers both linear and modular scalability, automatic failover support, consistent reads and writes, and automatic table sharding. In addition, users get a Java API for client access, a block cache for real-time queries, and an extensible JRuby-based (JIRB) shell.</p> <p>Couchbase Server is a multipurpose, distributed NoSQL database ideal for AI-powered applications and applications where data must be managed for user profiles and dynamic product catalogs. The database applies object modeling, provides flexible JavaScript Object Notation (<a href="https://www.theserverside.com/definition/JSON-Javascript-Object-Notation">JSON</a>) documents and supports distributed ACID transactions for NoSQL applications. Global cross-data center replication (XDCR) ensures database availability across <a href="https://www.techtarget.com/searchdatamanagement/feature/Should-you-host-your-database-on-site-or-in-the-cloud">clouds and on-premises</a> locations. Developers can write SQL queries for querying and transacting JSON data using Couchbase's AI-powered copilot called <i>Capella iQ</i>. Couchbase also provides many other features like vector and text search to build better user experiences, key-value access for improved agility and flexibility, multi-dimensional scaling (MDS) to independently scale database services in a cluster, and <a href="https://www.techtarget.com/searchsecurity/definition/role-based-access-control-RBAC">role-based access control</a> for enterprise-grade security.</p> <p>Amazon SimpleDB is a NoSQL distributed data store that simplifies data storage and access. Users need not manage database administration tasks like infrastructure provisioning, schema management and performance tuning. SimpleDB works with other Amazon web services like Amazon EC2 and <a href="https://www.techtarget.com/searchaws/definition/Amazon-Simple-Storage-Service-Amazon-S3">Amazon S3</a>. Ideal for uses such as online games and indexing Amazon S3 object metadata, SimpleDB creates multiple geographically distributed data replicas to provide high data availability, durability, and flexibility.</p> <p>FoundationDB is an open source <a href="https://www.techtarget.com/searchdatamanagement/definition/multimodel-database">multimodel database</a> that can store and distribute many different data types safely. All the data is replicated in the database's key-value store component. It acts as an ACID database and features a distributed, scalable, fault-tolerant architecture. This database is suitable for applications that need to support heavy loads without appreciably increasing costs.</p> <p><i>Certain initiatives require specific considerations when choosing database software. For instance, with IoT initiatives, SQL vs. NoSQL is an issue, as is static vs. streaming. Find out what to assess when </i><a href="https://www.techtarget.com/iotagenda/tip/How-to-select-the-right-IoT-database-architecture"><i>selecting a database for an IoT project</i></a><i>.</i></p> </section> A distributed database is a database that consists of two or more files located in different sites on the same or different networks. https://cdn.ttgtmedia.com/visuals/digdeeper/6.jpg https://www.techtarget.com/searchoracle/definition/distributed-database Tue, 22 Jul 2025 15:43:00 GMT <p>MySQL is a popular, scalable, user-friendly, open source and free relational database management system (<a href="https://www.techtarget.com/searchdatamanagement/definition/RDBMS-relational-database-management-system">RDBMS</a>) that uses Structured Query Language (<a href="https://www.techtarget.com/searchdatamanagement/definition/SQL">SQL</a>) to store, manage and manipulate data.</p> <p>Well-known for its high speed, performance and reliability, MySQL is operating system (<a href="https://www.techtarget.com/whatis/definition/operating-system-OS">OS</a>)-agnostic so it runs seamlessly on numerous platforms, including <a href="https://www.techtarget.com/searchdatacenter/definition/Linux-operating-system">Linux</a>, <a href="https://www.techtarget.com/searchdatacenter/definition/Unix">UNIX</a>, <a href="https://www.techtarget.com/whatis/definition/Mac-OS">macOS</a> and <a href="https://www.techtarget.com/searchwindowsserver/definition/Windows">Windows</a>. MySQL can be used to support a wide range of applications, but it is most often associated with web applications and online publishing (websites and blogs).</p> <section class="section main-article-chapter" data-menu-title="What is MySQL used for?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What is MySQL used for?</h2> <p>Today, MySQL is a very popular RDBMS choice with both web developers and web-based organizations. Many of the top websites in the world and countless corporate and consumer-facing web applications use MySQL, including:</p> <ul class="default-list"> <li><a href="https://www.techtarget.com/whatis/definition/Facebook">Facebook</a>.</li> <li>X (formerly <a href="https://www.techtarget.com/whatis/definition/Twitter">Twitter</a>).</li> <li>YouTube.</li> <li><a href="https://www.techtarget.com/searchcio/definition/sharing-economy">Airbnb</a>.</li> <li><a href="https://www.techtarget.com/whatis/definition/Uber">Uber</a>.</li> <li><a href="https://www.techtarget.com/searchitoperations/definition/GitHub">GitHub</a></li> <li>Booking.com.</li> </ul> <p>MySQL is a particularly popular choice with websites using the .<a href="https://www.techtarget.com/whatis/definition/com">com</a> domain.</p> <p>Many content management systems (<a href="https://www.techtarget.com/searchcontentmanagement/definition/content-management-system-CMS">CMS</a>) are also based on MySQL, including the following:</p> <ul class="default-list"> <li>WordPress.com.</li> <li>WordPress.org.</li> <li>Drupal.</li> <li>Joomla.</li> <li>Contao.</li> <li>TYPO3.</li> <li>MODx.</li> </ul> <p>Other web applications and sites that use MySQL include the following:</p> <ul class="default-list"> <li>URL shortening and link management services like <a href="https://www.techtarget.com/whatis/definition/Bitly">Bitly</a>.</li> <li>Demand-side platforms (DSP) like Zemanta.</li> <li><a href="https://www.techtarget.com/searchmobilecomputing/definition/microblogging">Microblogging</a> platforms like Tumblr.</li> <li>Online avatar creation sites like Gravatar.</li> <li>Forum creation software like phpBB and MyBB.</li> <li><a href="https://www.techtarget.com/searchcio/definition/e-commerce">E-commerce</a> platforms like Shopify, WooCommerce and BigCommerce.</li> </ul> <p>MySQL is also commonly used as an embedded database for many software and hardware products.</p> </section> <section class="section main-article-chapter" data-menu-title="How MySQL works"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How MySQL works</h2> <p>MySQL is based on a <a href="https://www.techtarget.com/searchnetworking/definition/client-server">client-server</a> model, with a <a href="https://www.techtarget.com/whatis/definition/multithreading">multithreaded</a> server supporting multiple clients and fulfilling requests from them. The core of MySQL is MySQL server, which handles all of the database instructions (or commands). MySQL server is used for storing, organizing and managing data. Users can interact with MySQL server to submit queries and find data. MySQL server is available as a separate (standalone) program for use in a client-server networked environment and as a multithreaded library that can be embedded (or linked) into separate applications.</p> <p>MySQL operates along with several utility programs that support the administration of MySQL databases. Commands are sent to MySQL server via the MySQL client, which is installed on a computer. It is the MySQL client that queries the MySQL server to find the specified data.</p> <p>MySQL was originally developed to handle large databases quickly. Although MySQL is typically installed on only one machine, it is able to send the database to multiple locations, and users are able to access it via different MySQL client graphical <a href="https://www.techtarget.com/searchapparchitecture/definition/user-interface-UI">user interfaces</a>. These interfaces send SQL statements or <i>requests</i> to the server, extract data from the database and then display the results.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/dbJMCmtLZuc?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> </section> <section class="section main-article-chapter" data-menu-title="Core MySQL features"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Core MySQL features</h2> <p>MySQL is a free and open source RDBMS, so anyone can modify and distribute it under the GNU General Public License (<a href="https://www.techtarget.com/searchdatacenter/definition/GNU-General-Public-License-GNU-GPL-or-simply-GPL">GPL</a>). It is a <a href="https://www.techtarget.com/searchdatamanagement/definition/relational-database">relational database</a>, meaning it organizes <a href="https://www.techtarget.com/whatis/definition/data-point">data points</a> with defined relationships and stores data in tables of rows and columns which are defined by <a href="https://www.techtarget.com/searchdatamanagement/definition/schema">schemas</a>. This structural approach makes it easy to organize and find different types of information in the database, such as text, numbers and dates.</p> <p>MySQL enables data to be stored and accessed across multiple storage engines, including InnoDB, cluster shared volumes (CSV) and NDB (NDB Cluster). This provides greater flexibility and freedom in use. MySQL is also capable of replicating data and partitioning tables for better performance and durability. MySQL users aren't required to learn new commands; they can access their data using standard SQL commands.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.theserverside.com/rms/onlineImages/server_side-mysql_innodb_vs_myisam-f.png"> <img data-src="https://www.theserverside.com/rms/onlineImages/server_side-mysql_innodb_vs_myisam-f_mobile.png" class="lazy" data-srcset="https://www.theserverside.com/rms/onlineImages/server_side-mysql_innodb_vs_myisam-f_mobile.png 960w,https://www.theserverside.com/rms/onlineImages/server_side-mysql_innodb_vs_myisam-f.png 1280w" alt="MySQL InnoDB vs. MyISAM comparison image" height="218" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>MySQL InnoDB and MyISAM are two variations of the open source database management system -- here's how they compare. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p>MySQL is written in C and C++ and is available across over 20 platforms, including Mac, Windows, Linux and Unix. The RDBMS supports large databases with millions of records and supports many <a href="https://www.techtarget.com/searchapparchitecture/definition/data-type">data types</a> including signed or unsigned integers 1, 2, 3, 4 and 8 bytes long; FLOAT; DOUBLE; CHAR; VARCHAR; BINARY; VARBINARY; TEXT; BLOB; DATE; TIME; DATETIME; TIMESTAMP; YEAR; SET; ENUM; and OpenGIS spatial types. Fixed- and variable-length string types are also supported.</p> <p>MySQL supports a wide variety of application programming interfaces (<a href="https://www.techtarget.com/searchapparchitecture/definition/application-program-interface-API">APIs</a>). The APIs can link data to different back ends, programs, libraries and administrative tools in different programming languages to the MySQL server. Supported languages include <a href="https://www.theserverside.com/definition/Java">Java</a>, <a href="https://www.techtarget.com/whatis/definition/Python">Python</a>, <a href="https://www.theserverside.com/definition/JavaScript">JavaScript</a>, <a href="https://www.techtarget.com/searchdatamanagement/definition/C">C++</a>, <a href="https://www.techtarget.com/whatis/definition/C-Sharp">C#</a> and <a href="https://www.techtarget.com/whatis/definition/PHP-Hypertext-Preprocessor">PHP</a>.</p> <p>For security, MySQL uses an access privilege and encrypted password system that enables host-based verification. MySQL clients can connect to MySQL server using several protocols, including <a href="https://www.techtarget.com/searchnetworking/definition/TCP-IP">TCP/IP</a> sockets on any platform. MySQL also supports a number of client and utility programs, command-line programs and administration tools such as MySQL Workbench. Other security features include data <a href="https://www.techtarget.com/searchsecurity/definition/encryption">encryption</a>, <a href="https://www.techtarget.com/searchsecurity/definition/data-masking">data masking</a>, secure storage and a database <a href="https://www.techtarget.com/searchsecurity/definition/firewall">firewall</a> to further strengthen data protection and also facilitate compliance with privacy laws and regulations like the General Data Protection Regulation (<a href="https://www.techtarget.com/whatis/definition/General-Data-Protection-Regulation-GDPR">GDPR</a>), Health Insurance Portability and Accountability Act (<a href="https://www.techtarget.com/searchhealthit/definition/HIPAA">HIPAA</a>) and Payment Card Industry Data Security Standard (<a href="https://www.techtarget.com/searchsecurity/definition/PCI-DSS-Payment-Card-Industry-Data-Security-Standard">PCI-DSS</a>).</p> <p>MySQL supports <a href="https://www.techtarget.com/searchdatamanagement/definition/ACID">ACID</a> (atomicity, consistency, isolation and durability) transactions to facilitate dependable and accurate processing of database transactions. These transactions also ensure that the database always remains in a consistent state, even in the event of a system failure.</p> <p>Another key feature of MySQL is its ability to scale. With built-in features like <a href="https://www.techtarget.com/searchstorage/definition/partition">partitioning</a> and clustering, MySQL can handle large-scale databases and high volumes of concurrent connections (high traffic applications). This is one reason why MySQL is a popular RDBMS choice for large organizations and for many web-based applications where fast, high-volume transactions are the norm, including social media, CMS and e-commerce.</p> </section> <section class="section main-article-chapter" data-menu-title="Advantages of MySQL"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Advantages of MySQL</h2> <p>One of the biggest advantages of MySQL over other RDBMSes is that it is easy to use. As a zero-administration database system, it is very easy to manage. It also installs very quickly, with a typical installation taking just a few minutes from start to end.</p> <p>MySQL is also a very reliable and scalable database. It is continuously maintained and improved by an active global community and has been tested for a wide variety of scenarios over many years. Also, it is based on a native replication architecture for data redundancy. As a result, MySQL offers consistently high performance for many kinds of web applications, including <a href="https://www.techtarget.com/searchitoperations/definition/mission-critical-computing">business-critical</a> and high-traffic applications.</p> <p>MySQL incorporates numerous native replication technologies to support <a href="https://www.techtarget.com/searchdatacenter/definition/high-availability">high availability</a>. Cloud-based MySQL implementations also include built-in <a href="https://www.techtarget.com/searchdisasterrecovery/definition/disaster-recovery">disaster recovery</a> features to boost <a href="https://www.techtarget.com/searchdisasterrecovery/definition/operational-resilience">operational resilience</a> and ensure high <a href="https://www.techtarget.com/whatis/definition/uptime-and-downtime">uptime</a>. Support for ACID transactions ensures data integrity, dependable transaction processing and consistent data modifications.</p> <p>Flexibility is another important benefit of MySQL. Developers can freely develop traditional SQL and NoSQL schema-free database applications, knowing that they will always work with MySQL. In addition, relational data and JavaScript Object Notation, or <a href="https://www.theserverside.com/definition/JSON-Javascript-Object-Notation">JSON</a>, documents can be combined within the same database.</p> <p>The numerous security features of MySQL help protect data within the database from misuse and unauthorized manipulation. Secure data storage and backup speed up disaster recovery and minimize the potential for data losses in the event of an adverse event.</p> <p>Finally, MySQL is a cost-effective RDBMS. As an open source product, anyone can download and use MySQL, often without incurring any cost. For example, the MySQL Enterprise Edition for Developers is completely free of cost for learning, developing and prototyping. That said, organizations looking to use MySQL for the purpose of embedding MySQL code into one or more commercial applications can purchase a commercially licensed version like MySQL Enterprise Edition from Oracle.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/businessanalytics-open_source_database_comparison-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/businessanalytics-open_source_database_comparison-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/businessanalytics-open_source_database_comparison-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/businessanalytics-open_source_database_comparison-f.png 1280w" alt="Open source database comparison image" height="392" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>In addition to MySQL, there are a number of other open source database systems available, each with its own strengths and best use cases. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="History of MySQL"> <h2 class="section-title"><i class="icon" data-icon="1"></i>History of MySQL</h2> <p>Originally conceived by the Swedish company MySQL AB in 1995, MySQL was meant to provide database users with a user-friendly RDBMS software. MySQL AB company was acquired by Sun Microsystems in 2008 and then by Oracle when it bought Sun in 2010. Oracle continues to own MySQL as of 2025.</p> <p>In 2000, MySQL was released under the GNU GPL, giving users the freedom to run, share, and even modify the software as per their needs. Today, MySQL remains open source and developers can still use it under the GNU GPL. However, for-profit enterprises must obtain a commercial license of MySQL from Oracle. As of 2025, MySQL is the most widely used open source RDBMS, accounting for the highest market share in the relational databases software market, well past many available alternatives like PostgreSQL and <a href="https://www.techtarget.com/searchaws/definition/Amazon-Relational-Database-Service-RDS">Amazon RDS</a>.</p> <p>Over the years, numerous offshoots of MySQL, also known as forks, have emerged, such as the following:</p> <ul class="default-list"> <li>MariaDB is a MySQL offshoot designed by the original developers of MySQL to be a drop-in replacement for MySQL offering features like clustering and multithreading for high availability, performance and scalability.</li> <li>Percona is an enterprise-grade open source software for MySQL, with built-in backup, observability and troubleshooting tools.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="MySQL editions"> <h2 class="section-title"><i class="icon" data-icon="1"></i>MySQL editions</h2> <p>MySQL is available in several different versions. Users can select the edition they need depending on their requirements and whether they are looking for a commercial (paid) or free product.</p> <h3>MySQL Enterprise Edition</h3> <p>This is the commercial version of MySQL. Suitable for developing, deploying and managing business-critical enterprise MySQL applications, it is highly scalable, secure and reliable, and provides very high levels of performance and uptime. Additionally, it is an ACID-compliant database.</p> <p>This edition includes numerous management tools and advanced features for MySQL, such as the ones listed below:</p> <ul class="default-list"> <li>MySQL Enterprise Backup for online data backup and recovery.</li> <li>MySQL Enterprise Authentication so organizations can use their existing security infrastructures.</li> <li>Security features like data encryption (Transparent Data Encryption and Enterprise Encryption), key generation and <a href="https://www.techtarget.com/searchsecurity/definition/digital-signature">digital signatures</a> to protect sensitive data.</li> <li>MySQL Enterprise Firewall that blocks database attacks such as <a href="https://www.techtarget.com/searchsoftwarequality/definition/SQL-injection">SQL injections</a>.</li> <li>Support for MySQL InnoDB Cluster and MySQL InnoDB ClusterSet to ensure high database availability.</li> <li>Built-in MySQL Enterprise Telemetry to monitor and track MySQL performance.</li> <li>MySQL Operator for <a href="https://www.techtarget.com/searchitoperations/definition/Google-Kubernetes">Kubernetes</a> to help businesses deploy and manage private DBaaS and microservices-based applications.</li> </ul> <p>MySQL Enterprise Edition also includes access to MySQL HeatWave, a cloud-based, fully managed database service that accelerates MySQL query performance. HeatWave combines transactional (<a href="https://www.techtarget.com/searchdatacenter/definition/OLTP">OLTP</a>) and analytical (<a href="https://www.techtarget.com/searchdatamanagement/definition/OLAP">OLAP</a>) workloads, supports <a href="https://www.techtarget.com/searchdatamanagement/definition/data-lakehouse">lakehouse</a>-scale analytics and provides integrated and automated <a href="https://www.techtarget.com/searchenterpriseai/definition/generative-AI">generative AI</a> capabilities that enable users to have contextual conversations with MySQL in natural, human-like language.</p> <p>MySQL Enterprise Edition also includes 24/7 support from Oracle. Oracle's support teams assist MySQL users in the development, deployment and management of MySQL applications.</p> <h3>MySQL Standard Edition</h3> <p>This edition delivers high-performance and scalable OLTP applications. It is a fully integrated, transaction-safe, ACID compliant database that allows organizations to minimize their database <a href="https://www.techtarget.com/searchdatacenter/definition/TCO">TCO</a>. This version incorporates MySQL Replication to ensure high availability and scalability. It also includes MySQL Workbench for database architects, developers and <a href="https://www.techtarget.com/searchdatamanagement/definition/database-administrator">database administrators</a>. MySQL Workbench is a unified visual environment for development and design. It also includes integrated administration tools for server configuration, user administration and data backup.</p> <h3>MySQL Classic Edition</h3> <p>This is another low-cost version of MySQL. This embedded, high performance and zero administration database also helps reduce database TCO, particularly for independent software vendors (<a href="https://www.techtarget.com/searchitchannel/definition/ISV">ISVs</a>), original equipment manufacturers (<a href="https://www.techtarget.com/searchitchannel/definition/OEM">OEMs</a>) and value-added resellers (<a href="https://www.techtarget.com/searchitchannel/definition/VAR">VARs</a>) developing read-intensive applications (using the MyISAM storage engine). As with the other editions of MySQL, MySQL Classic Edition supports 20+ platforms and operating systems, including Linux, Unix, Mac and Windows for greater development and deployment flexibility.</p> <h3>MySQL Enterprise Edition for Developers</h3> <p>This is the free version of MySQL. It is meant for developers who want to learn, develop and prototype MySQL.</p> <p>This edition includes all the features of MySQL Enterprise Edition:</p> <ul class="default-list"> <li>A fast, multithreaded, multiuser SQL database server.</li> <li>A C++ interface for communicating with MySQL servers.</li> <li>An online "hot" backup solution with data compression technology to protect data and ensure high performance.</li> <li>A driver to implement the Java Database Connectivity (<a href="https://www.theserverside.com/definition/Java-Database-Connectivity-JDBC">JDBC</a>) API.</li> <li>A fully-managed ADO.NET driver for MySQL (Windows only).</li> <li>A standardized database driver for Windows, Linux, Mac OS X and Unix platforms.</li> <li>A lightweight <a href="https://www.techtarget.com/searchapparchitecture/definition/middleware">middleware</a> that provides transparent routing between applications and backend MySQL Servers.</li> <li>An interactive JavaScript/Python/SQL interface to perform data queries and administration operations.</li> </ul> <p>Developers can download MySQL Enterprise Edition for Developers directly from Oracle.</p> <p>Oracle also offers the following:</p> <ul class="default-list"> <li><b>MySQL NDB Cluster CGE</b> is a distributed database that<b> </b>provides linear scalability, high availability (HA), in-memory real-time access and transactional consistency for mission critical applications.</li> <li><b>MySQL as an Embedded Database</b> is an embedded database for applications, hardware and appliances, and is meant for use by ISVs, OEMs and VARs.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="MySQL and LAMP"> <h2 class="section-title"><i class="icon" data-icon="1"></i>MySQL and LAMP</h2> <p>MySQL is an important component of an open source enterprise stack called <a href="https://www.techtarget.com/whatis/definition/LAMP-Linux-Apache-MySQL-PHP">LAMP</a>. LAMP is a web development platform that stands for Linux, Apache, MySQL and PHP/Python/Perl. Thus, it uses Linux as the OS, <a href="https://www.techtarget.com/whatis/definition/Apache">Apache</a> as the web server, MySQL as the RDBMS and PHP (or <a href="https://www.techtarget.com/whatis/definition/Perl">Perl</a> or Python instead of PHP) as the <a href="https://www.techtarget.com/searchapparchitecture/definition/object-oriented-programming-OOP">object-oriented</a> scripting language.</p> <p>Many applications running on the LAMP stack rely on MySQL to store and manage data. These include numerous websites (static and dynamic) and high-performance web applications. The LAMP combination is popular because it is tried-and-tested and therefore does not require rigorous testing of libraries, modules and tools. It also offers flexibility and ease of use for a variety of web applications. Moreover, the four technologies are open source, freely available for anyone to use and maintained by a large and highly active IT community. All four LAMP technologies also support many other free and open source tools and frameworks that developers can use to create and host web content.</p> </section> <section class="section main-article-chapter" data-menu-title="MySQL vs. PostgreSQL"> <h2 class="section-title"><i class="icon" data-icon="1"></i>MySQL vs. PostgreSQL</h2> <p>PostgreSQL is another popular open source relational database. In active development since 1986, PostgreSQL is feature-rich, reliable and extensible. In fact, PostgreSQL offers more features than MySQL, along with greater flexibility in data types and concurrency.</p> <p>One major difference between PostgreSQL and MySQL is that PostgreSQL is an object-relational database management system (ORDBMS). This means that it has an object-oriented design that allows programmers to communicate with database servers using objects in code and also define complex custom data types and functions. In contrast, MySQL is a relational database management system that stores data in rows and columns and defines relationships between data points.</p> <p>PostGreSQL is suitable as a general-purpose OLTP database, particularly for mission-critical applications and high-traffic websites. It is also ideal for dynamic websites and applications where customer data, transactions, and product catalogs need to be reliably and securely managed such as for e-commerce platforms. PostgreSQL is also suitable as a geospatial database (when used with the PostGIS extension) and as a federated database for various use cases.</p> <p>Like MySQL, PostgreSQL runs on all major operating systems and is also ACID-compliant. In addition, it includes numerous security features to protect data integrity and securely manage data sets of any size. Both MySQL and PostgreSQL are open source and supported by a dynamic community of users and developers.</p> </section> <section class="section main-article-chapter" data-menu-title="Compatibility with other services"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Compatibility with other services</h2> <p>MySQL was designed to be compatible with other systems. It supports deployment in virtualized environments, such as Amazon RDS for MySQL, Amazon RDS for MariaDB and Amazon Aurora for MySQL. Users can transfer their data to a Microsoft SQL Server database by using database migration tools like AWS Schema Conversion Tool and the AWS Database Migration Service.</p> <p>However, when migrating from Microsoft SQL Server to MySQL, the architectural differences must be considered. The database object <a href="https://www.techtarget.com/searchdatamanagement/definition/semantic-technology">semantics</a> between Microsoft SQL Server and MySQL are similar, but not identical. In MySQL, there is no difference between a database and a schema, so the terms are used interchangeably, while SQL Server treats the two as separate entities.</p> <p><i>Open source databases are strong alternatives to proprietary ones. Here are a <a href="https://www.techtarget.com/searchdatamanagement/feature/Top-open-source-databases-to-consider">dozen options to consider</a></i><i>.</i></p> </section> MySQL is a popular, scalable, user-friendly, open source and free relational database management system (RDBMS) that uses Structured Query Language (SQL) to store, manage, and manipulate data. https://cdn.ttgtmedia.com/visuals/digdeeper/5.jpg https://www.techtarget.com/searchoracle/definition/MySQL Fri, 18 Jul 2025 14:33:00 GMT <p>Pattern recognition is the ability to detect arrangements of characteristics in data that yields information about a given system or <a href="https://www.techtarget.com/whatis/definition/data-set">data set</a>. Pattern recognition is a main ability of machine learning (<a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-ML">ML</a>), <a href="https://www.techtarget.com/searchdatamanagement/definition/data-analytics">data analytics</a> and <a href="https://www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence">AI</a>.</p> <p>In a technological context, a pattern might be recurring sequences or aspects in data that can be used to make a prediction. These patterns might be simple relationships between <a href="https://www.techtarget.com/whatis/definition/variable">variables</a>, or they might be complex multifaceted relationships that use several variables. Pattern recognition is using computers and algorithms to find and make predictions based on the hidden relationships within data.</p> <p>A simple example of pattern recognition would be an <a href="https://www.techtarget.com/whatis/definition/algorithm">algorithm</a> that tries to determine a person's gender based on their shopping history. It might classify someone who buys makeup and skirts as female, and someone who buys sports memorabilia and work boots as male. This prediction might be then fed into another system to make recommendations for future purchases.</p> <p>If the process leads to a wrong prediction or it classifies something the wrong way, it is known as a <a href="https://www.techtarget.com/whatis/definition/prediction-error">prediction error</a>.</p> <section class="section main-article-chapter" data-menu-title="Where is pattern recognition used?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Where is pattern recognition used?</h2> <p>Pattern recognition is one of the most powerful abilities of mathematics and computers. It has found applications in almost every area of modern science, technology and industry.</p> <p>Over time it can be used to predict trends, find particular configurations of features in images that identify objects, detect frequent combinations of words and phrases for natural language processing (<a href="https://www.techtarget.com/searchbusinessanalytics/definition/natural-language-processing-NLP">NLP</a>), or locate particular clusters of behavior on a network that could indicate an attack -- among almost endless other possibilities.</p> <p>Pattern recognition is essential to many overlapping areas of IT, including <a href="https://www.techtarget.com/searchbusinessanalytics/definition/big-data-analytics">big data analytics</a>, <a href="https://www.techtarget.com/searchsecurity/definition/biometric-authentication">biometric authentication</a>, security and artificial intelligence.</p> <p>Some examples of pattern recognition include the following:</p> <ul class="default-list"> <li><a href="https://www.techtarget.com/searchenterpriseai/definition/facial-recognition">Facial recognition</a> software takes in data related to the characteristics of a person's face and uses an algorithm to match that specific pattern to an individual record in a database.</li> <li>Pattern recognition algorithms in meteorological software can detect recurring relationships among weather data that can be used to forecast probable future weather events.</li> <li>Intrusion detection systems (<a href="https://www.techtarget.com/searchsecurity/definition/intrusion-detection-system">IDS</a>) and <a href="https://www.techtarget.com/whatis/definition/threat-intelligence-cyber-threat-intelligence">threat intelligence</a> hunting software rules describe patterns of behaviors and events that can indicate illegitimate network traffic.</li> <li><a href="https://www.techtarget.com/searchenterpriseai/definition/machine-vision-computer-vision">Machine vision</a> uses pattern recognition to classify the input from a camera. A simple system for industrial use might be trained for a very specific purpose, such as determining if an apple is ripe or not. A more complex system might be trained to detect and classify many objects in a live video feed.</li> <li>Accounting and financial systems might use pattern recognition to predict sales data or forecast expenses. Other systems might be able to detect fraudulent activity on an account.</li> <li>Large language models (<a href="https://www.techtarget.com/whatis/definition/large-language-model-LLM">LLM</a>) use pattern recognition at their core. They use incredibly large data sets to find the hidden relationships among words and use them to make a prediction of what word to use next based on a given input.</li> </ul> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineImages/enterpriseai-machine_learning_models_cheat_sheet-f.png"> <img data-src="https://www.techtarget.com/rms/onlineImages/enterpriseai-machine_learning_models_cheat_sheet-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineImages/enterpriseai-machine_learning_models_cheat_sheet-f_mobile.png 960w,https://www.techtarget.com/rms/onlineImages/enterpriseai-machine_learning_models_cheat_sheet-f.png 1280w" alt="Machine learning models and their training algorithms cheat sheet." height="380" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Pattern recognition is at the heart of machine learning models that use supervised, unsupervised, semisupervised and reinforcement learning methods. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="How does pattern recognition work?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How does pattern recognition work?</h2> <p>There are myriad methods of pattern recognition, but they can be broadly classified as either <a href="https://www.techtarget.com/searchenterpriseai/definition/supervised-learning">supervised learning</a> or <a href="https://www.techtarget.com/searchenterpriseai/definition/unsupervised-learning">unsupervised learning</a>.</p> <p>With supervised learning prelabeled and structured data is used to try to produce specific outputs. In unsupervised learning, the training data is unorganized and the system is used to identify important characteristics.</p> <p>Some methods use bits of both approaches, called semisupervised learning. In <a href="https://www.techtarget.com/searchenterpriseai/definition/reinforcement-learning">reinforcement learning</a>, the model is further refined as it receives user feedback when it produces good or bad output.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/rHeaoaiBM6Y?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <p>Several steps are typically used while developing a pattern recognition system:</p> <ol class="default-list"> <li><b>Data collection.</b> Information is gathered, sorted and preprocessed.</li> <li><b>Feature extraction.</b> Important features of the data are identified and then used to make predictions.</li> <li><b>Classification.</b> The extracted features are analyzed to find the needed relationships and make predictions.</li> </ol> <p>Further post-processing and refinement stages might be applied.</p> <p>Statistical methods use mathematics to detect and classify the input. For numeric data, this can be simple mathematical regressions or linear mathematics. Other data might be converted into <a href="https://www.techtarget.com/whatis/definition/vector">vectors</a>, which can be operated on mathematically.</p> <p><a href="https://www.techtarget.com/searchenterpriseai/definition/neural-network">Neural networks</a> do pattern recognition by simulating a neural pathway. The decision-making is stored as the relationship between the nodes in the network.</p> <p>In clustering methods, similarities are found in certain attributes that can be used to identify subgroups.</p> <p>In syntactic methods, the relationships and hierarchy among the data are used. These can be rules or <a href="https://www.techtarget.com/searchenterpriseai/definition/decision-tree-in-machine-learning">decision trees</a>. This is similar to how the syntax in a language describes how words are grouped.</p> <p>In <a href="https://www.techtarget.com/searchenterpriseai/definition/fuzzy-logic">fuzzy</a> systems, a level of uncertainty is kept in the process. The output might allow for multiple or incomplete matches. This can help in situations where the data is unclear or organic in nature.</p> <p><a href="https://www.techtarget.com/whatis/definition/template">Template</a> matching is the simplest method. Rules for what attributes correspond to what output are developed and applied.</p> <p><i>AI is similar to human intelligence, but there are important differences between them. Here are </i><a href="https://www.techtarget.com/searchenterpriseai/tip/Artificial-intelligence-vs-human-intelligence-How-are-they-different"><i>some key ways that AI and human thinking differ</i></a><i>.</i></p> </section> Pattern recognition is the ability to detect arrangements of characteristics in data that yields information about a given system or data set. https://cdn.ttgtmedia.com/visuals/digdeeper/2.jpg https://www.techtarget.com/whatis/definition/pattern-recognition Thu, 17 Jul 2025 06:13:00 GMT <p>Consulting and systems integration companies are among the hundreds of launch partners backing the new AWS agentic AI cloud marketplace, which the hyperscaler rolled out today.</p> <p>Partners such as Accenture, Cognizant, Deloitte, Presidio, PwC, Tata Consultancy Services and Wipro provide offerings on AWS' AI Agents and Tools marketplace, which functions within the flagship AWS Marketplace. <a href="https://www.techtarget.com/searchenterpriseai/definition/agentic-AI">Agentic AI</a> products and services have been available on the AWS Marketplace, but the new agentic AI one differs as a distinctly branded category and for its integration with Amazon Bedrock AgentCore services. <a href="https://www.techtarget.com/searchenterpriseai/news/366627853/AWS-launches-AgentCore-system-and-agentic-marketplace">AgentCore</a>, also announced today, lets customers deploy and run AI agents.</p> <p>AWS' agentic AI marketplace move tracks with the broader trend toward fulfilling customer demand for <a href="https://www.techtarget.com/searchitchannel/news/366627020/Partners-begin-to-target-agentic-AI-marketplace-platforms">agentic AI in marketplaces</a>, where IT service providers play a growing role. Google Cloud Marketplace, Microsoft's Azure Marketplace and Salesforce's AppExchange also offer subsections that focus on agentic AI.</p> <p>Chris Sullivan, vice president of AWS channels and alliances for the Americas, cited an "incredible increase" in demand and opportunity for agentic AI as influencing the timing of the new marketplace's launch.</p> <p>To meet that demand, Sullivan anticipates that partners will be able to build complex offerings that combine professional services and software in the AI Agents and Tools marketplace. In the AWS Marketplace, partners often use the platform's Channel Partner Private Offer, or CPPO, feature to bundle and <a href="https://www.techtarget.com/searchitchannel/news/366623839/AWS-Marketplace-channel-partners-rev-software-service-sales">sell software and services</a>.</p> <p>AWS specifically cites professional services as one of the types of "agent solutions" available in the AI Agents and Tools marketplace. The <a target="_blank" href="https://aws.amazon.com/marketplace/solutions/ai-agents-and-tools/" rel="noopener">company said</a> customers will be able to "develop and implement an AI strategy with specialized professional services from AWS Partners." The other categories include embedded agents, prebuilt agents, agent tools and agent development offerings.</p> <p><i>John Moore is a writer for Informa TechTarget covering the CIO role, economic trends and the IT services industry.</i></p> Accenture, Cognizant and Deloitte are among the partners with offerings on AWS' AI Agents and Tools marketplace, which is expected to combine professional services and software. https://cdn.ttgtmedia.com/rms/onlineimages/iot_g1182604383.jpg https://www.techtarget.com/searchitchannel/news/366627818/Consultants-SIs-back-AWS-agentic-AI-marketplace-launch Wed, 16 Jul 2025 18:09:00 GMT <p>A prediction error is the failure of a model of a system to accurately forecast outcomes. It is the difference between the predicted and measured value. The mean absolute error, mean squared error and mean squared predictive error are among several methods used to calculate accuracy of the <a href="https://www.techtarget.com/searchenterpriseai/definition/predictive-modeling">predictive model</a>.</p> <p>Errors are an inevitable element of predictive analytics that should be quantified and presented along with any model, often in the form of a confidence interval that indicates the expected accuracy of predictions. Analysis of prediction errors from similar or previous models can help determine confidence intervals.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/w5CculFJ718?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <p>It is impossible to eliminate prediction errors. Instead, the goal is to minimize and classify the error. When the potential error is understood, an analyst can have confidence in the values given and make decisions based on the predicted values.</p> <p>Predictive analytics software processes new and historical data to forecast activity, behavior and trends. The programs apply <a href="https://www.techtarget.com/whatis/definition/statistical-analysis">statistical analysis </a>techniques, analytical queries and machine learning algorithms to data sets to create predictive models that quantify the likelihood of a particular event happening.</p> <section class="section main-article-chapter" data-menu-title="Types of prediction errors"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Types of prediction errors</h2> <p>The underlying training data is often the source of prediction errors in machine learning (<a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-ML">ML</a>) models. Here are some common problems:</p> <ul class="default-list"> <li><b>Bias.</b> <a href="https://www.techtarget.com/searchenterpriseai/feature/Big-data-throws-big-biases-into-machine-learning-data-sets">Bias</a> is a given input that is over- or undervalued compared to an objective value.</li> <li><b>Underfitting.</b> The training model is not complex enough to capture the important characteristics of the data, resulting in overgeneralized guesses as the output.</li> <li><b>Overfitting</b><b>.</b> With <a href="https://www.techtarget.com/whatis/definition/overfitting-in-machine-learning">overfitting</a>, the training model is too complex and captures noise or unwanted features of the training data, resulting in an overly sensitive output.</li> <li><b>Variance.</b> Small changes in the training data cause large changes in the final model.</li> <li><b>Generalization error.</b> The model makes poor predictions when given novel data outside the training set.</li> </ul> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineImages/bi_ezine-how_ai_systems_amplify_bias.png"> <img data-src="https://www.techtarget.com/rms/onlineImages/bi_ezine-how_ai_systems_amplify_bias_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineImages/bi_ezine-how_ai_systems_amplify_bias_mobile.png 960w,https://www.techtarget.com/rms/onlineImages/bi_ezine-how_ai_systems_amplify_bias.png 1280w" alt="An infographic that provides an example -- involving image recognition -- of the problem of bias in AI data." height="336" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Image recognition is just one area where AI systems can amplify bias. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Prediction errors in AI and ML"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Prediction errors in AI and ML</h2> <p>In <a href="https://www.techtarget.com/searchenterpriseai/Ultimate-guide-to-artificial-intelligence-in-the-enterprise">AI</a>, the analysis of prediction errors can help guide ML, similarly to the way it does for human learning.</p> <p>In <a href="https://www.techtarget.com/searchenterpriseai/definition/reinforcement-learning">reinforcement learning</a>, for example, an <a href="https://www.techtarget.com/searchenterpriseai/definition/agent-intelligent-agent">agent</a> might use the goal of minimizing error feedback as a way to improve. Prediction errors, in that case, might be assigned a negative value and predicted outcomes a positive value, in which case the AI is programmed to attempt to maximize its score. That approach to ML, sometimes called <i>error-driven learning</i>, seeks to stimulate learning by approximating the human drive for mastery.</p> <p>ML has many more potential uses than statistical analysis. ML tools might be used in speech recognition or <a href="https://www.techtarget.com/searchenterpriseai/definition/machine-vision-computer-vision">machine vision</a> tasks. In such tasks, the parameters and features of the data and model can be examined to determine why the model misclassified a given input.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/components_of_a_machine_vision_system-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/components_of_a_machine_vision_system-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/components_of_a_machine_vision_system-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/components_of_a_machine_vision_system-f.png 1280w" alt="Components of a machine vision diagram." height="262" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Cameras and sensors are used to collect images as light or photons and turn them into electric signals or electrons for processing and use in industrial applications. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p>When <a href="https://www.techtarget.com/searchenterpriseai/tip/AI-model-optimization-How-to-do-it-and-why-it-matters">building a model</a> on table data, it can be difficult to determine what caused the error. The exact correlation between the values and why a specific output is given may not be intuitive.</p> </section> <section class="section main-article-chapter" data-menu-title="How to minimize prediction errors"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How to minimize prediction errors</h2> <p>Human analysis of prediction errors is crucial. When predictions fail, humans can use metacognitive functions, examining prior predictions and failures and deciding, for example, whether there are <a href="https://www.techtarget.com/whatis/definition/correlation">correlations</a> and trends, such as consistently being unable to foresee outcomes accurately in particular situations.</p> <p>Applying that type of knowledge can inform decisions and improve the quality of future predictions. These additional steps can help minimize prediction errors:</p> <ol class="default-list"> <li><b>Data sanitization</b><b>.</b> Inspect source data for outliers, incompleteness and inaccuracies.</li> <li><b>Data transformation.</b> Normalize and format the data.</li> <li><b>Parameter tuning.</b> Select and test different model training parameters.</li> <li><b>Regularization.</b> Use L1 or L2 regularization to penalize errors and avoid overfitting.</li> <li><b>Prediction testing.</b> Analyze model outputs in a controlled setting, and refine as needed.</li> </ol> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/regularization_in_machine_learning-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/regularization_in_machine_learning-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/regularization_in_machine_learning-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/regularization_in_machine_learning-f.png 1280w" alt="An infographic that explains how regularization can help alleviate overfitting and underfitting problems in machine learning." height="392" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Regularization is an important technique that can alleviate overfitting and underfitting problems in machine learning. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p>The performance of any predictive model should be evaluated continually. Are the given outputs over an acceptable error threshold? Has the input data changed from when the model was trained? Is the model still meeting key success criteria? Iterating and continually refreshing the model can help to minimize prediction errors.</p> <p><em>Regularization in ML represents a valuable set of techniques that can help mitigate the risk of overfitting. By employing these strategies, data scientists can enhance the performance of ML models and reduce potentially costly errors, ultimately benefiting the overall outcome of their work. Learn more about <a href="https://www.techtarget.com/searchenterpriseai/feature/Machine-learning-regularization-explained-with-examples">ML regularization with examples</a>.</em></p> </section> A prediction error is the failure of a model of a system to accurately forecast outcomes. https://cdn.ttgtmedia.com/visuals/digdeeper/6.jpg https://www.techtarget.com/whatis/definition/prediction-error Tue, 24 Jun 2025 09:00:00 GMT <p>Descriptive analytics is a type of data analytics that looks at past data to give an account of what has happened. Results are typically presented in reports, dashboards, bar charts and other visualizations that are easily understood.</p> <p>The field of <a href="https://www.techtarget.com/searchdatamanagement/definition/data-analytics">data analytics</a> is generally divided into four main types: descriptive analytics, diagnostic analytics, <a href="https://www.techtarget.com/searchbusinessanalytics/definition/predictive-analytics">predictive analytics</a> and prescriptive analytics. A fifth type, real-time analytics, <a href="https://www.techtarget.com/searchbusinessanalytics/feature/7-enterprise-use-cases-for-real-time-streaming-analytics">analyzes data as it's generated, collected or updated</a>.</p> <p>Descriptive analytics is the simplest of these techniques. It can be used by itself or treated as a preliminary stage of data processing to create a summary or abstraction that, in turn, supports further investigation, analysis or actions performed by other types of analytics.</p> <section class="section main-article-chapter" data-menu-title="How does descriptive analytics work?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How does descriptive analytics work?</h2> <p>Descriptive analytics uses various <a href="https://www.techtarget.com/whatis/definition/statistical-analysis">statistical analysis</a> techniques to slice and dice raw data into a form that enables people to see patterns, identify anomalies, improve planning and compare things. Enterprises realize the most value from descriptive analytics when using it to compare items over time or against each other. For example, a finance manager might compare product sales month over month or against related categories.</p> <p>Descriptive analytics can work with numerical data, qualitative data or some combination of the two. Numerical data might quantify things like revenue, profit or a physical change. Qualitative data might characterize elements such as gender, ethnicity, profession or political party. To improve understanding, raw numerical data is often binned into ranges or categories such as age ranges, income brackets or zip codes.</p> <p>Descriptive analysis techniques perform various mathematical calculations that make recognizing or communicating a pattern of interest easier. For example, <i>central tendency</i> describes what is normal for a given data set by considering characteristics such as the average, mean and median. Other elements include frequency, variation, ranking, range and deviation.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/5_modes_of_analytics-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/5_modes_of_analytics-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/5_modes_of_analytics-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/5_modes_of_analytics-f.png 1280w" alt="Diagram showing five modes of analytics" height="286" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>The five modes of analytics help organizations understand what and why something happened, what could happen in the future and what should be done next. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Real examples of descriptive analytics"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Real examples of descriptive analytics</h2> <p>Descriptive analytics is used in a variety of industries. Some real examples include the following:</p> <ul class="default-list"> <li><b>Financial statement analysis.</b> Organizations use descriptive analytics to evaluate performance based on an analytical review of their data, such as income statements, balance sheets and cash flow reports. Gartner <a href="https://www.gartner.com/en/finance/trends/autonomous-finance-predictions" target="_blank" rel="noopener">predicts</a> that at least 90% of descriptive analytics done in finance will be fully automated by 2027.</li> <li><b>Retail sales reporting.</b> Descriptive analytics is used by businesses to track sales data across different periods of time. This can give companies an insight into purchasing trends, product demand and seasonal variations.</li> <li><b>Healthcare records.</b> Descriptive analytics is important in hospitals and other health clinics, as it enables healthcare providers to identify health conditions, treatment outcomes and resource utilization.</li> <li><b>Website traffic analytics.</b> Digital marketers use descriptive analytics to track website visitor trends, bounce rates and engagement levels.</li> <li><b>Social media engagement analysis.</b> Businesses use descriptive analytics to measure likes, shares, comments and follower growth, which can help them adjust content strategies and improve audience interaction.</li> </ul> <p>Descriptive analytics is also commonly used for the following:</p> <ul class="default-list"> <li>Planning a new program.</li> <li>Measuring the effectiveness of a new program.</li> <li>Understanding sales trends.</li> <li>Comparing companies.</li> <li>Motivating behavior with KPIs.</li> <li>Recognizing anomalous behavior.</li> <li>Interpreting survey results.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="What can descriptive analytics tell you?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What can descriptive analytics tell you?</h2> <p>Businesses use descriptive analytics to assess, compare, spot anomalies and identify relative strengths and weaknesses. Let's walk through how these might work in practice.</p> <ul class="default-list"> <li><b>Assess. </b>Developing a description of the various aspects of business operations forms a baseline for a company and can be used to help assess behavior that improves or weakens outcomes. Important metrics include business performance (sales, costs and profits), website performance (click-through rate, conversion rate) and <a href="https://www.techtarget.com/whatis/definition/customer-satisfaction-CSAT">customer satisfaction</a> (<a href="https://www.techtarget.com/searchcustomerexperience/definition/Net-Promoter-Score-NPS">Net Promoter Score</a>, customer satisfaction score and social media likes) or equipment performance (uptime, repair time and maintenance cost).</li> <li><b>Compare. </b>Baseline metrics for a characteristic can be compared across time, categories, programs or interventions. Existing sales performance could be compared to last month's sales, to the same period a year ago or over a more extended range to understand historical trends. Similarly, teams might compare sales across product categories or compare one company's performance against a related company.</li> <li><b>Spot anomalies. </b>Teasing apart descriptive statistics can sometimes reveal outliers worthy of further investigation. For example, descriptive analytics might show that sales in one region are significantly higher than others, or that a successful product line suddenly slumps. These anomalies could prompt additional research using diagnostic analytics to understand their root causes.</li> <li><b>Identify relative strengths and weaknesses. </b>A company might discover, for example, that it tends to sell better to younger shoppers than to older ones. This insight might inspire efforts to double down on strengths, with more focused marketing or prompt action to address weaknesses in the area where sales are lagging.</li> </ul> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/LdXtLWdeS68?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> </section> <section class="section main-article-chapter" data-menu-title="Steps in descriptive analytics"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Steps in descriptive analytics</h2> <p>The descriptive analytics process includes the following steps:</p> <ol class="default-list"> <li><b>Quantify goals</b><b>.</b> The process starts by translating some broad business goals, such as better business performance, into specific, measurable outcomes such as sales per product, <a href="https://www.techtarget.com/whatis/definition/cost-per-sale-CPS">cost per sale</a> or conversion rate.</li> <li><b>Identify relevant data</b><b>.</b> Teams need to identify any types of data that could help improve the understanding of the critical metric. The data might be buried across one or more internal systems or within various third-party data sources.</li> <li><b>Organize data</b><b>.</b> Data from different sources, applications or teams needs to be cleaned and normalized to improve analytics accuracy.</li> <li><b>Analysis</b><b>.</b> Various statistical and mathematical techniques combine, summarize and compare the raw data in different ways to generate data features.</li> <li><b>Presentation</b><b>.</b> Data features might be numerically presented in a report, dashboard or visualization. Common <a href="https://www.techtarget.com/searchbusinessanalytics/tip/12-data-visualization-techniques-for-effective-BI-applications">visualization techniques</a> include bar charts, pie charts, line charts, bubble charts and histograms.</li> </ol> </section> <section class="section main-article-chapter" data-menu-title="Benefits and drawbacks of descriptive analytics"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Benefits and drawbacks of descriptive analytics</h2> <p>The use of descriptive analytics can provide the following benefits:</p> <ul class="default-list"> <li>It can simplify communication about numerical data.</li> <li>It can improve understanding of complex situations.</li> <li>Companies can compare performance against the competition or across product lines.</li> <li>It can help motivate teams to reach new goals.</li> </ul> <p>Drawbacks and weaknesses of descriptive analytics include the following:</p> <ul class="default-list"> <li>Existing biases can be amplified either accidentally or deliberately.</li> <li>Results can direct a company's focus to metrics that are not helpful, such as sales versus profits.</li> <li>Motivational metrics can be gamed to encourage unintended behavior, such as mouse movers or sales fraud.</li> <li>Poorly chosen metrics can lead to a false sense of security.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="Descriptive analytics tools"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Descriptive analytics tools</h2> <p>Relatively simple tools, like an Excel spreadsheet, and some knowledge of business management are enough to craft basic descriptive analytics. However, teams might see greater value across the organization by scaling various types of tools to democratize analytics development and promote the sharing of business intelligence (<a href="https://www.techtarget.com/searchbusinessanalytics/definition/business-intelligence-BI">BI</a>).</p> <p>Business intelligence tools like <a href="https://www.techtarget.com/searchcontentmanagement/definition/Microsoft-Power-BI">Power BI</a>, <a href="https://www.techtarget.com/whatis/definition/Tableau">Tableau</a> and <a href="https://www.techtarget.com/searchbusinessanalytics/definition/Qlik">Qlik</a> can simplify many steps of the descriptive analytics process.</p> <p>Descriptive analytics tools provide various ways for reorganizing raw data to view new patterns by calculating characteristics such as averages, frequencies, variations, rankings, ranges and deviations. While these basic techniques are baked into essential BI tools, a team might turn to more sophisticated data science tools for complex statistics, including the following:</p> <ul class="default-list"> <li>The <a href="https://www.techtarget.com/searchbusinessanalytics/definition/R-programming-language">R programming language</a>.</li> <li>IBM's <a href="https://www.techtarget.com/whatis/definition/SPSS-Statistical-Package-for-the-Social-Sciences">Statistical Package for the Social Sciences</a>.</li> <li>Analytics software from <a href="https://www.techtarget.com/searchbusinessanalytics/definition/SAS-Institute-Inc">SAS Institute Inc</a>.</li> <li><a href="https://www.techtarget.com/searchbusinessanalytics/news/252492435/KNIME-focused-on-performance-speed-of-analytics-platform">Knime open source software</a>.</li> <li>SQL.</li> <li>Python.</li> </ul> <p>Data-wrangling tools can help automate data engineering processes by cleansing, reformatting and combining data from many different sources. Popular tools include offerings from Alteryx, Cambridge Semantics, Trifacta, Talend and Tamr.</p> </section> <section class="section main-article-chapter" data-menu-title="Descriptive analytics vs. prescriptive, predictive and diagnostic analytics"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Descriptive analytics vs. prescriptive, predictive and diagnostic analytics</h2> <p>As noted, the field of analytics is commonly characterized as including four main kinds of capabilities:</p> <ul class="default-list"> <li><b>Descriptive analytics</b>, as explained, provides information about what happened. You might see, for example, an increase in sales following a new promotion.</li> <li><b>Diagnostic analytics</b> examines data more closely to understand the causes of events and behaviors. For example, in the case of increased sales, you might investigate which categories of people showed the greatest response and why this might be the case.</li> <li><b>Predictive analytics</b> identifies future probabilities and trends based on a model of past behavior. For example, once you have identified the root cause of that uptick in sales, predictive analytics could help calculate the likelihood and magnitude of a similar sales increase happening in other markets.</li> <li><b>Prescriptive analytics</b> makes recommendations or automates decisions based on a given prediction. For example, <a href="https://www.techtarget.com/searchcio/definition/Prescriptive-analytics">prescriptive analytics</a> could suggest the best ways to structure and implement a successful sales promotion in another region based on that region's local demographics.</li> </ul> <div class="youtube-iframe-container"> <iframe id="ytplayer-1" src="https://www.youtube.com/embed/NSe1JBETWJs?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> </section> <section class="section main-article-chapter" data-menu-title="Descriptive analytics vs real-time analytics"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Descriptive analytics vs real-time analytics</h2> <p>Descriptive analytics and <a href="https://www.techtarget.com/searchcustomerexperience/definition/real-time-analytics">real-time analytics</a> are both critical methods of data analysis. While descriptive analytics focuses on analyzing past data to identify trends, patterns and insights, real-time analytics processes incoming data. Therefore, organizations use these separate methods to make different kinds of decisions.</p> <p>Descriptive analytics, for instance, gives organizations insight into long-term customer behavior and trends, which makes it critical for research-based strategic planning. On the other hand, real-time analytics' continually updated data sets are best suited for immediate decisions, in which quick responses are needed. These include fraud detection, dynamic traffic management and live monitoring of financial transactions.</p> </section> <section class="section main-article-chapter" data-menu-title="Jobs using descriptive analytics"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Jobs using descriptive analytics</h2> <p>Descriptive analytics is critical to a variety of jobs across industries. Some of the most commons jobs include the following:</p> <ul class="default-list"> <li><b>Data analyst.</b> Analyzes structured datasets to identify trends, patterns and KPIs, helping optimize business strategies.</li> <li><b>Business intelligence analyst.</b> Uses data visualization and reporting tools to provide insights for executives and managers.</li> <li><b>Marketing analyst.</b> Conducts research on customer behavior, campaign effectiveness and market trends using descriptive analytics.</li> <li><b>Financial analyst.</b> Evaluates past financial performance using statistical methods to assess profitability, investment risks and budget planning.</li> <li><b>Healthcare data analyst.</b> An increasingly important position that works with <a href="https://www.techtarget.com/healthtechanalytics/feature/The-Healthcare-Data-Cycle-Generation-Collection-and-Processing">patient records and clinical data sets</a> to improve healthcare services, operational efficiency and patient outcomes.</li> </ul> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/business_intelligence_analyst_vs_data_analyst-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/business_intelligence_analyst_vs_data_analyst-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/business_intelligence_analyst_vs_data_analyst-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/business_intelligence_analyst_vs_data_analyst-f.png 1280w" alt="Business intelligence analyst vs. data analyst checklist" height="385" width="559"> <figcaption> <i class="icon pictures" data-icon="z"></i>Business intelligence analysts and data analysts are essentially differentiated by their main goals: The former focus on analyzing internal business trends, while the latter have a broader scope. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p><i>Descriptive analytics is an important method used in data analysis. Learn about </i><a href="https://www.techtarget.com/searchbusinessanalytics/feature/8-types-of-bias-in-data-analysis-and-how-to-avoid-them"><i>nine types of bias in data analysis you should avoid</i></a><i>. Further explore the differences of </i><a href="https://www.techtarget.com/searchbusinessanalytics/tip/Descriptive-vs-prescriptive-vs-predictive-analytics-explained"><i>descriptive vs. prescriptive vs. predictive analytics</i></a><i>.</i></p> </section> Descriptive analytics is a type of data analytics that looks at past data to give an account of what has happened. https://cdn.ttgtmedia.com/visuals/digdeeper/5.jpg https://www.techtarget.com/whatis/definition/descriptive-analytics Mon, 02 Jun 2025 20:17:00 GMT <p>AI democratization, or the democratization of AI, is the concept that the benefits and control of <a href="https://www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence">artificial intelligence</a> should be fairly divided among all people. AI democratization efforts claim that the power of AI will transform many aspects of our daily lives and, if it is properly directed, these transformations can benefit the majority of people instead of only a select few.</p> <p>AI democratization is a new and changing area of discussion. It has started to solidify around four main areas, though: the wide availability of AI use, AI development, AI benefits and <a href="https://www.techtarget.com/searchenterpriseai/definition/AI-governance">AI governance</a>. These areas are not definitive and sometimes blur into one another.</p> <p>In addition to larger socioeconomic issues, how to better democratize AI within an organization is a common concern.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/WP6z_X5d-Rw?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <section class="section main-article-chapter" data-menu-title="Democratizing AI use"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Democratizing AI use</h2> <p>Democratizing AI use is the claim that artificial intelligence should be available to all regardless of nationality or economic status.</p> <p>Historically, new technological advances have taken a trickle-down path in adoption. Often, technology is first used by wealthy people in advanced economic countries, then as it becomes cheaper to produce or easier to use, it becomes available to a larger and larger population.</p> <p><a href="https://www.techtarget.com/searchenterpriseai/definition/generative-AI">Generative AI</a> can avoid this long process, as a truly <a href="https://www.techtarget.com/whatis/definition/disruptive-technology">disruptive innovation</a>. It is extremely easy to use, even a person who is unfamiliar with technology can adapt to using these tools quickly. Users only need a cellphone and internet connection to access them. Once deployed, it is relatively easy to add more users. With all this in place, AI could be used across the world.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/generative_ai_benefits_for_business-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/generative_ai_benefits_for_business-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/generative_ai_benefits_for_business-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/generative_ai_benefits_for_business-f.png 1280w" alt="Generative AI benefits for business diagram." height="498" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Generative AI is disruptive due to its ease of use and availability to anyone, no matter their expertise in AI. It is an example of AI democratization. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Democratizing AI development"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Democratizing AI development</h2> <p>Democratizing AI development is the principle that anyone should be able to work with AI and customize it for their needs. AI development has typically demanded a lot of resources, including expert-level knowledge, computing power and money. AI democratization involves facilitating development by providing user-friendly resources and support -- such as prebuilt <a href="https://www.techtarget.com/whatis/definition/algorithm">algorithms</a>, intuitive interfaces and high-performance <a href="https://www.techtarget.com/searchcloudcomputing/definition/cloud-computing">cloud computing</a> platforms. Having those supports in place makes it feasible for in-house developers without special expertise to create their own <a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-ML">machine learning</a> applications and other AI software.</p> <p>It is hoped that democratizing AI development can help alleviate issues with <a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-bias-algorithm-bias-or-AI-bias">AI bias</a>. AI's output is highly influenced by its training. Having a wider variety of people and viewpoints involved during training can, hopefully, make the resulting AI fairer and more equitable.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-1" src="https://www.youtube.com/embed/4qSZEP5lJi4?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> </section> <section class="section main-article-chapter" data-menu-title="Democratizing AI benefits"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Democratizing AI benefits</h2> <p>AI is a powerful tool. Like all new tools, it can make people's lives easier and allow them to accomplish more work with less effort. It is hoped that democratizing the benefits of AI will help most people instead of only a select few. This means that AI needs to be available regardless of socioeconomic factors or nationality, and that the efficiencies benefit workers as well as corporations.</p> <p>As AI continues to advance, it can widen the gap between people and countries that use it and those that do not. As certain jobs become more automated or even outright replaced by AI, how can those affected by these changes be helped instead of hurt?</p> <p>As a brief thought experiment, if AI use means that a certain employee can now do the same amount of work in half the time, how should this affect them? The same amount of work for the same amount of pay, but half the working hours, or twice the work for the same pay and same hours?</p> </section> <section class="section main-article-chapter" data-menu-title="Democratizing AI governance"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Democratizing AI governance</h2> <p><a href="https://www.techtarget.com/searchdatamanagement/tip/AI-data-governance-is-a-requirement-not-a-luxury">AI governance are the rules and standards</a> that AI must follow. This can include legal regulations, AI safeguards, AI bias, privacy and transparency. By democratizing AI governance, more people will have a say in these important aspects.</p> <p>Currently, the companies that develop and deploy AI have the most control over these aspects. Many argue that those affected by AI use should have the most say in AI governance.</p> <p>AI governance touches on areas like data protection laws, fair use and copyright. Existing legal frameworks in these areas might not be enough to cover the effects AI is having on society.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/trustworthy_ai_framework-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/trustworthy_ai_framework-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/trustworthy_ai_framework-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/trustworthy_ai_framework-f.png 1280w" alt="Trustworthy AI framework inforgraphic." height="493" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Possible frameworks for responsible AI can differ from company to company. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Democratizing AI in an organization"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Democratizing AI in an organization</h2> <p>Many organizations are grappling with similar issues in democratizing AI with their workers. They want all to be able to use and benefit from AI. This might necessitate AI training or adoption programs so that no one is left behind.</p> <p>By making AI development tools simple and available, workers can customize AI to their use. Also, by engaging with the workers, organizations can try to minimize bias or risks in the AI they use.</p> <p><i>What happens when we use AI outside of a small group of experts? To avoid business problems, leaders need to invest wisely in AI tools and training for their employees. Learn more about how the </i><a href="https://www.techtarget.com/searchenterpriseai/feature/Democratization-of-AI-creates-benefits-and-challenges"><i>democratization of AI creates benefits and challenges</i></a><i>. Also, explore the <a href="https://www.computerweekly.com/opinion/AI-governance-mapping-the-road-ahead">road ahead for AI governance</a></i><i>, see about <a href="https://www.techtarget.com/searchenterpriseai/feature/Differentiating-between-good-and-bad-AI-bias">differentiating between good and bad AI bias</a></i><i> and learn about the <a href="https://www.techtarget.com/whatis/feature/Pros-and-cons-of-AI-generated-content">pros and cons of AI-generated content</a></i><i>.</i></p> </section> AI democratization, or the democratization of AI, is the concept that the benefits and control of artificial intelligence should be fairly divided among all people. https://cdn.ttgtmedia.com/visuals/digdeeper/3.jpg https://www.techtarget.com/whatis/definition/AI-democratization Mon, 19 May 2025 20:35:00 GMT <p>Amazon Q Developer can offer users automated and optimized support for their software development needs.</p> <p>While it cannot fully replace the technical skills required by software developers to deliver code, it can optimize the quality and efficiency of the entire process. Like other AI tools, it is important to evaluate the suggestions and ensure they deliver the expected results.</p> <p>Learn more about Amazon Q Developer, how it can benefit development workloads and how to get started.</p> <section class="section main-article-chapter" data-menu-title="What is Amazon Q Developer?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What is Amazon Q Developer?</h2> <p>Amazon Q Developer, part of the Amazon Q portfolio, focuses on software development tasks, such as code generation, analysis and optimization. It also provides recommendations about AWS infrastructure, such as documentation, optimizations, best practices and troubleshooting for cloud resource configurations. Through a conversational interface, users can type prompts and define instructions on the tasks or analyses to execute.</p> <p>There are many ways for developers to interact with Amazon Q Developer, including the following:</p> <ul class="default-list"> <li>AWS console.</li> <li>Mobile app.</li> <li>Documentation pages.</li> <li>Chat applications.</li> <li>Integrated development environments (IDEs).</li> <li>CLIs.</li> </ul> <p>It can be accessed for application software development tasks from the UIs of AWS tools that handle code or SQL statements, such as Amazon EMR Studio, Amazon SageMaker AI Studio, JupyterLab, AWS Lambda, AWS Glue Studio, AWS Cloud9, <a href="https://www.techtarget.com/searchcloudcomputing/answer/Compare-EMR-Redshift-and-Athena-for-data-analysis-on-AWS">Amazon Athena and Amazon Redshift</a>. It also integrates with external Git repositories, like GitLab.</p> <p>For development and operational tasks, integrating it with your preferred IDE and using it through the AWS console or CLI are highly recommended.</p> <h3>AWS Console-to-Code feature</h3> <p>For some services, such as EC2, VPC and Relational Database Service, the console supports the AWS Console-to-Code feature. This enables Amazon Q Developer to record manual interactions in the console and convert them to code syntax for the AWS CLI, Cloud Development Kit and CloudFormation. Though this feature is limited to the mentioned services, it is expected to expand to other AWS resources.</p> </section> <section class="section main-article-chapter" data-menu-title="How IDE integration with Amazon Q Developer works"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How IDE integration with Amazon Q Developer works</h2> <p>With Amazon Q Developer's focus on software development tasks, its integration with IDE tools is particularly relevant. It supports commonly used IDEs, such as Visual Studio Code (VS Code), Eclipse, Visual Studio, JetBrains and the AWS coding environments mentioned above. For supported IDEs, download an Amazon Q plugin or extension for the specific IDE, which is available on their respective marketplace websites.</p> <p>Here is an example of how IDEs work with Amazon Q Developer. In this case, I used VS Code. After downloading and installing the extension for VS Code, the IDE requires users to log in to AWS.</p> <p>Keep in mind: The user must have sufficient Identity and Access Management (IAM) permissions. These can be provided by attaching the AmazonQDeveloperAccess managed policy to the IAM entity using the console. This gives the developer full access without administrator access. Then, invoke Amazon Q from the AWS console by clicking on the icon located at the top-right corner, which enables a chat interface.</p> <p>For the free tier of Amazon Q Developer, users must have an AWS Builder ID. For those using the monthly Pro subscription, they need access to an identity managed by AWS IAM Identity Center.</p> <p>Users should familiarize themselves with the tab for Amazon Q Developer.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/marquez_q_developer_howto_fig1-h.jpg"> <img data-src="https://www.techtarget.com/rms/onlineimages/marquez_q_developer_howto_fig1-h_mobile.jpg" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/marquez_q_developer_howto_fig1-h_mobile.jpg 960w,https://www.techtarget.com/rms/onlineimages/marquez_q_developer_howto_fig1-h.jpg 1280w" alt="The Amazon Q Developer tab."> <figcaption> <i class="icon pictures" data-icon="z"></i>The IDE displays a separate tab for Amazon Q Developer. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p>Developers can also select one or more lines of code and right-click. The extension displays several options. Each delivers a way to understand and optimize existing code. Consider the following examples:</p> <ul class="default-list"> <li><b>Explain.</b> Details the functionality delivered by a block of code, which can help clarify complex lines of code.</li> <li><b>Refactor.</b> Suggests recommendations that can deliver a better code structure or align with best practices.</li> <li><b>Fix.</b> If there are syntax errors in a piece of code, returns code snippets with the issue fixed.</li> <li><b>Optimize.</b> Provides recommendations for better performance or compute resource utilization and explains the reasoning.</li> <li><b>Generate tests.</b> Generates unit test files.</li> <li><b>Send to prompt.</b> Has the ability to start prompts with /dev, /test, /review, /doc or /transform, depending on the task to be performed.</li> <li><b>Inline chat. </b>Supports a chat interface that delivers these features using natural language capabilities.</li> </ul> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/marquez_q_developer_howto_fig2-h.jpg"> <img data-src="https://www.techtarget.com/rms/onlineimages/marquez_q_developer_howto_fig2-h_mobile.jpg" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/marquez_q_developer_howto_fig2-h_mobile.jpg 960w,https://www.techtarget.com/rms/onlineimages/marquez_q_developer_howto_fig2-h.jpg 1280w" alt="Amazon Q Developer code optimization options."> <figcaption> <i class="icon pictures" data-icon="z"></i>Use these helpful options within Amazon Q Developer to further optimize code. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Security"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Security</h2> <p>AWS emphasizes that the data made accessible to Amazon Q is not used to train its underlying models. This means sensitive data is securely kept inside organizations' AWS accounts and is not shared outside. This provides a secure environment for analysis, development, troubleshooting and operational tasks.</p> <p>Amazon Q is built on Amazon Bedrock, which has an automated abuse detection feature. This feature detects functionality that violates the <a href="https://aws.amazon.com/ai/responsible-ai/policy/" target="_blank" rel="noopener">AWS Responsible AI Policy</a>, ensuring that this tool cannot be used to create or optimize code intended for harmful or illegal purposes.</p> </section> <section class="section main-article-chapter" data-menu-title="Amazon Q Developer pricing"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Amazon Q Developer pricing</h2> <p>There is a free tier and a Pro subscription package, which charges a $19 monthly fee per subscribed user. For <a href="https://www.techtarget.com/searchcloudcomputing/tip/Implement-these-4-AWS-Organizations-best-practices">AWS Organizations</a>, each Pro user can have access to all AWS member accounts with a single monthly fee.</p> <p>A few features are unavailable in the free tier, but it provides similar functionality to the Pro subscription. Depending on the feature, users could see a limited number of interactions per month. These include the following:</p> <ul class="default-list"> <li>Generate or refine code.</li> <li>Integrate data.</li> <li>Generate or analyze SQL statements.</li> <li>Diagnose errors in the console.</li> </ul> <p>In general, the free tier is a good way to learn about the features available in Amazon Q Developer.</p> <p><i>Ernesto Marquez is owner and project director at Concurrency Labs, where he helps startups launch and grow their applications on AWS. He enjoys building serverless architectures, building data analytics solutions, implementing automation and helping customers cut their AWS costs.</i></p> </section> Amazon Q Developer is a versatile tool for software development, offering code generation, optimization recommendations and seamless integration with various IDEs and AWS tools. https://cdn.ttgtmedia.com/rms/onlineimages/maze_g1261411056.jpg https://www.techtarget.com/searchcloudcomputing/tip/Get-started-with-Amazon-Q-Developer Mon, 19 May 2025 09:00:00 GMT <p>Access to a wide range of system metrics is essential for operating reliable and scalable applications in the cloud. For Amazon Relational Database Service, users need system-level visibility to monitor performance and diagnose potential issues.</p> <p><a href="https://www.techtarget.com/searchaws/definition/CloudWatch">Amazon CloudWatch</a> provides detailed visibility into multiple metrics related to resources launched in AWS, including database clusters and instances managed in RDS. Even though CloudWatch metrics are a critical area to monitor, they don't offer much visibility into OS-level activity. However, Enhanced Monitoring in Amazon RDS could provide these deeper insights.</p> <p>RDS Enhanced Monitoring publishes metrics gathered by an agent running on the OS on which a particular database instance is launched. This additional set of data provides detailed metrics for database deployments managed by RDS that can be useful for the following:</p> <ul class="default-list"> <li>Early issue detection.</li> <li>Enhanced troubleshooting.</li> <li>Performance improvements.</li> </ul> <p>They can also be used to find optimal cloud resource allocations, which can reduce cost. The feature is available for all RDS database engines: MariaDB, MySQL, PostgreSQL, Db2, Oracle and Microsoft SQL Server.</p> <section class="section main-article-chapter" data-menu-title="How Enhanced Monitoring gathers metrics"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How Enhanced Monitoring gathers metrics</h2> <p>One important consideration regarding Enhanced Monitoring is that the OS data is not published as standard CloudWatch metrics. Instead, OS metrics are published into <a href="https://www.techtarget.com/searchcloudcomputing/tip/When-and-how-to-search-with-Amazon-CloudWatch-Logs">CloudWatch Logs</a> -- by default, into the RDSOSMetrics log group. Users must analyze OS log data and act upon it using a different approach compared to other RDS CloudWatch metrics.</p> <p><a href="https://www.techtarget.com/searchcloudcomputing/tip/Follow-these-examples-to-use-CloudWatch-Logs-Insights">CloudWatch Logs Insights</a> is a useful tool to query data stored in CloudWatch Logs. It uses a query syntax that supports specific data filtering and aggregation based on specific fields. CloudWatch metric filters can detect patterns in CloudWatch Logs and convert them into CloudWatch metrics. This can trigger anomaly detection, alarm notifications and actionable events based on OS log data. For simplified analysis and troubleshooting, admins can visualize these custom metrics in CloudWatch dashboards. For more detailed analysis and actions, use CloudWatch Logs subscription filters to automatically export incoming log data to other AWS services, such as Kinesis, Data Firehose and <a href="https://www.techtarget.com/searchaws/definition/AWS-Lambda-Amazon-Web-Services-Lambda">Lambda</a>.</p> <p>With Enhanced Monitoring, a unique log stream in the RDSOSMetrics log group ingests each database instance or cluster. This enables the separation of log data based on the source database, which is useful for data filtering when using the CloudWatch Logs console. The RDS console also supports a view of Enhanced Monitoring data using graph widgets, which simplifies the analysis of OS data within the context of specific RDS instances.</p> <h3>Overview of metrics</h3> <p>Metrics are published to CloudWatch Logs in JSON format, and each record contains source data identifiers. These include the associated RDS instance, database engine, record timestamp, number of virtual CPUs and amount of time the database has been active. Then, there are over 80 detailed metrics related to the following:</p> <ul class="default-list"> <li>CPU utilization.</li> <li>Disk utilization.</li> <li>File system.</li> <li>Process load.</li> <li>Memory utilization.</li> <li>Network utilization.</li> <li>Resource consumption.</li> </ul> <p>The same set of metrics is available for all database engines, except for Microsoft SQL Server. This engine supports a slightly narrower range of metrics.</p> </section> <section class="section main-article-chapter" data-menu-title="Enabling RDS Enhanced Monitoring"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Enabling RDS Enhanced Monitoring</h2> <p>Admins can turn on Enhanced Monitoring with the AWS SDK, CLI, CloudFormation or console. This is done at the cluster or instance level, depending on the RDS deployment configuration. If users enable a cluster, they must configure it at the cluster or instance level. Configure it by setting a value for the <span style="font-family: courier new, courier, monospace;">MonitoringInterval</span> parameter -- e.g., 0, 1, 5, 10, 15, 30 or 60 seconds. When this parameter is set to zero, Enhanced Monitoring is disabled.</p> <p>Specify <span style="font-family: courier new, courier, monospace;">MonitoringRoleArn</span>, which is an <a href="https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_Monitoring.OS.Enabling.html" target="_blank" rel="noopener">Identity and access Management role</a> that grants log ingestion permissions to RDS. These parameters can be configured during cluster/instance launch or as an update operation to an existing RDS resource.</p> </section> <section class="section main-article-chapter" data-menu-title="RDS Enhanced Monitoring pricing"> <h2 class="section-title"><i class="icon" data-icon="1"></i>RDS Enhanced Monitoring pricing</h2> <p>The Enhanced Monitoring feature itself does not incur any RDS cost. However, given that data is published to CloudWatch Logs, there is data ingestion and storage usage incurred in CloudWatch.</p> <p>Data ingestion costs $0.50 per gigabyte in the N. Virginia region, and 1 TB of data storage costs $4.50 per month. There is an OS metrics granularity configuration in RDS that supports log ingestion every 1, 5, 10, 15, 30 and 60 seconds. The lower the granularity, the higher the data ingestion and storage cost is in CloudWatch Logs.</p> <p>According to RDS, log data published every second -- the highest available frequency -- incurs approximately 16 GB per month, which costs about $8 in data ingestion fees. Users can also configure CloudWatch Logs retention to expire records after a period of time or to keep them indefinitely. Depending on application needs, the right balance between ingestion granularity and log retention can optimize cost.</p> <p><i>Ernesto Marquez is owner and project director at Concurrency Labs, where he helps startups launch and grow their applications on AWS. He enjoys building serverless architectures, building data analytics solutions, implementing automation and helping customers cut their AWS costs.</i></p> </section> RDS Enhanced Monitoring provides teams with additional data visibility to improve database scalability, performance, availability and cost optimizations. https://cdn.ttgtmedia.com/rms/onlineimages/pharma_g1358852671.jpg https://www.techtarget.com/searchcloudcomputing/tip/Gain-insights-with-Enhanced-Monitoring-in-Amazon-RDS Wed, 14 May 2025 09:00:00 GMT <p>Amazon Lightsail is designed for speed and simplicity. The instances run on top of Amazon EC2 and are bundled with other AWS resources, though those services are abstracted so they're not visible to the user.</p> <p>Simplicity comes with tradeoffs, however. Amazon EC2 enables far more ways to build, deploy and manage applications. It also remains the dominant AWS offering, alongside Amazon S3. But, for a subset of users, Lightsail is the better fit.</p> <p>Let's take a closer look at the Lightsail vs. EC2 debate to see how IT teams could benefit from Lightsail and to evaluate pricing.</p> <section class="section main-article-chapter" data-menu-title="Benefits of Amazon Lightsail"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Benefits of Amazon Lightsail</h2> <p><a href="https://www.techtarget.com/searchaws/definition/Amazon-Lightsail">Amazon Lightsail</a> is a virtual private server (VPS) that bundles compute, storage, networking and DNS. It also has built-in capabilities that include a managed database, load balancer, support for containers and content delivery network (<a href="https://www.techtarget.com/searchnetworking/definition/CDN-content-delivery-network">CDN</a>).</p> <p>Amazon Lightsail is for businesses that want to spin up a server without working through all the pricing, configuration and management details associated with a typical AWS deployment. Benefits organizations could see with Lightsail include the following:</p> <ul class="default-list"> <li>Standardized applications and configurations.</li> <li>Basic templates.</li> <li>Affordability.</li> <li>Developer-friendly features.</li> <li>Built-in security.</li> <li>Simplified UX.</li> </ul> <h3>Standardized applications and configurations</h3> <p>Developers can use Lightsail to build simple projects, such as a blog, website or basic e-commerce application, using standard applications and configurations. For example, to set up and configure a WordPress blog, just select a platform and a blueprint, such as a preconfigured WordPress instance. The following diagram illustrates how media content is delivered within Lightsail using its load balancer, database and S3 for WordPress.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/WP4_1.jpg"> <img data-src="https://www.techtarget.com/rms/onlineimages/WP4_1_mobile.jpg" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/WP4_1_mobile.jpg 960w,https://www.techtarget.com/rms/onlineimages/WP4_1.jpg 1280w" alt="Visualization of a load balancer" height="248" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>The load balancer works to deliver media content for a WordPress site. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <p>To do the same thing in EC2, users need to provision the instance, add <a href="https://www.techtarget.com/searchaws/tip/Troubleshoot-common-Amazon-EBS-performance-issues">Amazon Elastic Block Store</a> (EBS) block storage or Amazon S3 object storage, provision the image, and then configure all the different resources and applications.</p> <h3>Basic templates</h3> <p>With Lightsail, users can create the same basic template in a few clicks. These templates also make Lightsail an attractive staging environment for testing new applications or features before deploying them on live instances.</p> <h3>Affordability</h3> <p>Lightsail presents a handful of options that are deployable at a predictable monthly price. However, the service is not ideal for applications that require a highly configurable environment or consistently high CPU performance, such as video encoding or analytics.</p> <h3>Developer-friendly features</h3> <p>Amazon Lightsail comes with a wide range of preconfigured software stacks for common use cases. It also supports a comprehensive API that developers can use to facilitate communication with other cloud services.</p> <p>An SSH terminal interface enables developers to kick off more complex configurations and make changes using a standard shell interface. This also enables them to preconfigure scripts for particular scenarios. With a Secure FTP application, developers can copy application data from a backup or retrieve data from other systems as required using a standard FTP application.</p> <h3>Built-in security</h3> <p>Lightsail supports the same encryption and <a href="https://www.techtarget.com/searchsecurity/definition/identity-access-management-IAM-system">identity and access management</a> as the rest of the AWS cloud. It's easy to add firewall rules to a Lightsail instance to control the traffic that connects with it. Lightsail templates are also preconfigured with security best practices. This reduces the risk of a developer accidentally leaving an S3 storage bucket open to hackers.</p> <p>Amazon also automatically updates all applications and OS blueprints with new security patches. These can be applied automatically when running databases during a preconfigured maintenance window. However, these are not applied to running OSes, applications and container instances. As a result, the instance must be periodically stopped and restarted with the latest blueprint to take advantage of the latest updates.</p> <h3>Simplified UX</h3> <p>The AWS interface enables users to quickly set up a new instance from a checklist of OS and application templates. These templates have back-end, network interface and storage configurations for applications like WordPress, Drupal and Django. Choose the apps, location and name. Then, click <b>Create</b>, and the new service spins up in minutes. This can save considerable time for users unfamiliar with the nuances of cloud configuration and storage setup required with EC2.</p> </section> <section class="section main-article-chapter" data-menu-title="Comparing Lightsail vs. EC2"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Comparing Lightsail vs. EC2</h2> <p>The packaged nature of Lightsail makes it difficult to compare directly with EC2. IT teams need to connect AWS' flagship compute service with other AWS offerings -- each with its own pricing structure -- to create a viable environment to build and deploy applications.</p> <p>For this comparison, some Lightsail features are shown and contrasted with what can be done natively in AWS when using EC2 as the linchpin. All pricing listed is based on the U.S. East Region.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/comparison_of_amazon_lightsail_vs_ec2-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/comparison_of_amazon_lightsail_vs_ec2-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/comparison_of_amazon_lightsail_vs_ec2-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/comparison_of_amazon_lightsail_vs_ec2-f.png 1280w" alt="Compare Amazon Lightsail vs. EC2" height="347" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>Amazon Lightsail and EC2 can meet different needs for different organizations. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> <h3>Compute and block storage</h3> <p>When it comes to compute options, there is no comparison. Lightsail has eight virtual server sizes; EC2 has more than 750. Lightsail tops out at 16 cores and 64 GB of memory; EC2 instances can reach 448 cores and 12,288 GB of memory.</p> <p>Again, the point of Lightsail is not endless customization. For granularity and a massive range of configuration options, go with EC2 instead.</p> <p>In terms of solid-state drive (SSD) storage, Lightsail's disk sizes range from 20 GB to 1,280 GB. EC2 has far more flexibility, but in most cases, users need to sort out the attached instance storage separately through EBS, which can cost an additional $0.08 per GB. With Lightsail, that is all preconfigured.</p> <p>Teams that outgrow their VPS instance or need more control can take a snapshot and export it to a new instance in EC2.</p> <h3>Containers</h3> <p>Developers can use Lightsail containers to start loading standard Docker or other container images into the cloud. However, this service lacks the <a href="https://www.techtarget.com/searchcloudcomputing/answer/Amazon-ECS-vs-Kubernetes-Which-should-you-use-on-AWS">fine-grained controls</a> of Amazon ECS or Amazon Elastic Kubernetes Service.</p> <p>Lightsail containers cost significantly more for the raw resources. The basic container service costs $7 per month for one-quarter of a virtual CPU (vCPU) with 512 MB of RAM and goes up to $160 per month for four vCPUs with 8 GB of RAM. An equivalent EC2 instance is $3.02 on the low end and $97 on the high end. The Lightsail packages also include data egress quotas, which are an additional cost with EC2.</p> <p>All container services come with a flat 500 GB per month of transfer quotas, which otherwise adds $45 to the EC2 equivalent. Although this might tip the balance in favor of Lightsail containers, the main benefit lies in easy experimentation with basic container principles rather than cost savings.</p> <h3>Load balancer</h3> <p>Lightsail load balancers distribute traffic across instances in different <a href="https://www.techtarget.com/searchcloudcomputing/tip/Understand-AWS-Regions-vs-Availability-Zones">availability zones</a> (AZs). This addresses scaling issues and improves performance and redundancy. The load balancer also handles certificate management.</p> <p>Lightsail's load balancers aren't designed to handle consistently high traffic volumes, however. AWS recommends developers use EC2 with Application Load Balancer instead for workloads that involve more than 5 GB of data per hour, 400,000 new connections per hour or 15,000 active connections running at the same time.</p> <p>The Lightsail load-balancing service is priced at $18 per month. With EC2, developers can pick from Application, Network, Gateway and Classic Load Balancers. These services are charged on a consumption basis, so costs depend on the amount of traffic processed.</p> <h3>Content delivery network</h3> <p>Developers can bundle a Lightsail CDN built on top of the Amazon CloudFront network to improve the performance of applications accessed around the world. This is useful for websites, such as a blog, in which assets like CSS style sheets, JavaScript code, graphics and videos can be staged closer to users to reduce the response times. This can also reduce the load on the server, which can improve overall performance.</p> <p>EC2 works directly with <a href="https://www.techtarget.com/searchcloudcomputing/answer/Cloudflare-vs-Amazon-CloudFront-Which-CDN-is-right-for-you">Amazon CloudFront or third-party CDN services</a>, and it's better suited for complex configurations or workloads that require a lot of requests or video streaming.</p> <p>The bottom tier of the Lightsail CDN provides 50 GB per month free for the first year and then $2.50 per month afterward. At the high end, 500 GB costs $35 per month. For Amazon CloudFront, traffic is charged based on data transfers out and HTTP requests. Data transfer pricing varies, with higher-volume tiers translating to lower per-gigabyte rates.</p> <h3>Managed databases</h3> <p>AWS supports managed databases in the standard and high availability Lightsail tiers. This simplifies the selection process for MySQL databases, as well as efforts to launch, secure, monitor and maintain those repositories.</p> <p>The service automatically maintains a seven-day rolling database backup. Developers can configure longer-duration backups with snapshots, which are billed separately. These bundles are more expensive than Linux or Windows instances and have much more limited transfer allowances.</p> <p>The high availability tier supports redundancy and failover to servers in different Amazon AZs. The standard managed databases start at $15 per month for 1 GB of RAM, 40 GB of storage and 100 GB of data transfers. These go up to $115 per month for 8 GB of RAM, 240 GB of storage and 200 GB of data transfers. The high availability tiers are priced at double these rates.</p> <p>Lightsail managed databases don't provide the same level of performance or throughput that larger databases, such as MongoDB or Cassandra, might require. EC2 instances with provisioned IOPS SSD storage are a better option than Lightsail in these cases.</p> <p>Lightsail can work with other AWS database offerings. It supports Amazon DynamoDB, <a href="https://www.techtarget.com/searchcloudcomputing/answer/When-should-I-use-Amazon-RDS-vs-Aurora-Serverless">Amazon Relational Database Service and Amazon Aurora</a>, but you might need to peer to a separate Amazon VPC to make it work.</p> <p>The peering technique can connect to many, but not all, of the other AWS services. Peering is not required for smaller subsets of services, including Amazon S3 and Amazon CloudFront.</p> <h3>Data science research</h3> <p>Amazon Lightsail for Research helps data scientists and AI developers quickly spin up an environment to analyze large data sets without configuring or managing the cloud infrastructure components. This reduces the need for a powerful local computer. Both non-GPU and GPU-based versions are available, which perform better on larger analytics projects.</p> <p>These environments come preconfigured with popular data analytics applications, such as JupyterLab, RStudio and Scilab. Additional packages and extensions for various data science workflows are available.</p> <p>Upload data sets from a web browser to get started, and then delete the instance when they are done. The service also has budgeting controls to ensure that a particular analysis does not exceed a preconfigured cost limit.</p> <p>There are no charges for uploading data, but there is a limit of 0.5 terabytes for standard tiers and 1 TB for GPU tiers for free data downloads per month, after which $0.09 per GB egress charges come into effect. This is the same for all Lightsail for Research instance sizes. In most analytics use cases, the resulting analysis or processed data tends to be smaller than the analyzed data.</p> <p>Lightsail for Research is priced on an hourly basis, unlike the rest of the offerings mentioned previously, which are priced monthly. The top three tiers all come with GPUs. The most expensive tier, at $3.18 per hour, supports 16 vCPUs and 64 GB of RAM. All the tiers support 50 GB of SSD storage. At the bottom, pricing starts at $0.90 per hour for a server with four vCPUs and 8 GB of RAM.</p> <p>At the end of an analytics cycle, Amazon stops the instance. However, it continues to charge $0.00685 per hour, which comes to $4.93 per month. This might be useful if a data scientist wants to save intermediate work for subsequent processing or make it easier to return to a particular set of configurations.</p> <p>It's not clear what EC2 instance the standard tiers are comparable with. However, at the high end, GPU 4XL is comparable to g4dn.4xlarge, which costs $1.204 per hour and comes with 225 GB of storage. Lightsail for Research probably costs more per hour, but it could be a better option for data scientists to set up experiments without relying on IT or development teams.</p> </section> <section class="section main-article-chapter" data-menu-title="Lightsail vs. EC2 pricing"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Lightsail vs. EC2 pricing</h2> <p>Amazon Lightsail <a target="_blank" href="https://aws.amazon.com/lightsail/pricing/" rel="noopener">costs</a> are lower for basic resource usage compared to the On-Demand Instances in EC2.</p> <p>Linux/Unix, the smallest Lightsail instance, with 512 MB of RAM and 20 GB of SSD storage, is $3.50 per month to deploy. This is comparable to the EC2 equivalent, t3.nano, priced at $0.0052 per hour or about $3.38 monthly. However, this doesn't include SSD or data transfer costs. With a similar amount of SSD storage to the Lightsail package, t3.nano comes to $4.98 per month.</p> <p>At the high end, the $380 per month Lightsail package includes 64 GB of RAM, 1,280 GB of SSD storage and 8 TB of data transfers. Compare this to its EC2 equivalent, m6g.4xlarge, which comes with 64 GB of RAM for $443.52 per month, or $5,545.92 per month with 640 GB of SSD storage.</p> <p>Windows Lightsail servers are priced higher than the Linux variants due to licensing fees. The Windows plans start at $8 per month for 512 MB of RAM, 30 GB of SSD and 1 TB of data transfers. The high-end Windows plan costs $570 per month and includes 64 GB of RAM, 1,280 GB of SSD storage and 8 TB of data transfers.</p> <h3>Transfers and other fees</h3> <p>The Lightsail packages allow 1 TB to 8 TB of free data transfers, depending on instance size. EC2 instances cost $0.09 per GB transferred after the first 100 GB. This could add up to an additional $81 for 1 TB and $711 for 8 TB of outbound data transferred per month.</p> <p>Data transfer savings represent one of the biggest cost differentiators between Lightsail and EC2. Note that the allowances are considerably smaller for other Lightsail services. However, it's challenging to dynamically scale Lightsail services up or down in response to load. This might result in higher costs to overprovision resources to prepare for spikes.</p> <h3>Remember to delete</h3> <p>In theory, a developer can start and stop a Lightsail instance and save money when it's not running. AWS still charges for Lightsail instances even when an instance is stopped. To suspend charges, a developer must back up the instance and delete it from Lightsail. The enterprise must also pay another fee to keep an IP address in use -- to help maintain web server continuity -- that is no longer associated with the Lightsail instance. This adds $.005 per hour or $3.60 per month.</p> <p>Each account is limited to 20 Lightsail instances, five static IP addresses and three DNS zones. That might be fine for simple use cases, but an enterprise is unlikely to build a large-scale Lightsail deployment within these confines.</p> <p>One major downside is that Lightsail does not dynamically scale as well as EC2. Any apparent cost savings might be spent on overprovisioning resources the team never consumes.</p> </section> <section class="section main-article-chapter" data-menu-title="Competitive services"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Competitive services</h2> <p>There are a variety of competitive VPS services, each with different competitive advantages. Consider DigitalOcean, which focuses on simplifying the developer experience for provisioning standard services as Droplets, thereby simplifying management. It also provides a developer-friendly UI to deploy a library of preconfigured instances. DigitalOcean is priced comparably to Lightsail.</p> <p>Companies using other AWS applications should consider Lightsail since it can simplify integration and billing across multiple services. Other services make more sense for companies first exploring or testing various applications.</p> <p><b>Editor's note: </b><i>This article was updated to include additional information on the differences between Amazon Lightsail and EC2.</i></p> <p><i>George Lawton is a journalist based in London. Over the last 30 years, he has written more than 3,000 stories about computers, communications, knowledge management, business, health and other areas that interest him.</i></p> </section> Developers can use Amazon Lightsail to quickly build websites and applications, but EC2 affords teams more customization. Which service best benefits your organization's needs? https://cdn.ttgtmedia.com/rms/onlineimages/competition_a110169470.jpg https://www.techtarget.com/searchcloudcomputing/tip/Compare-Amazon-Lightsail-vs-EC2-for-your-web-app-needs Fri, 02 May 2025 09:00:00 GMT <p>Amazon (Amazon.com) is the world's largest online retailer and one of the largest providers of cloud services. As of 2025, it is considered a giant in both e-commerce and cloud computing.</p> <p>Headquartered in Seattle, Amazon has individual websites, software development centers, customer service centers, <a href="https://www.techtarget.com/searchdatacenter/definition/data-center">data centers</a> and fulfillment centers around the world. The company was founded by Jeff Bezos in 1994; he remained its CEO and president until 2021. As of 2025, he remains Amazon's executive chair and one of its largest individual shareholders.</p> <section class="section main-article-chapter" data-menu-title="What is Amazon known for?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What is Amazon known for?</h2> <p>Originally an online bookselling company, Amazon has morphed into an internet-based business enterprise that provides cloud computing, digital streaming and <a href="https://www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence">AI</a> services, in addition to an <a href="https://www.techtarget.com/searchcio/definition/e-commerce">e-commerce</a> platform.</p> <p>The <a href="https://www.techtarget.com/searchitoperations/definition/platform">platform</a> operates in more than 20 countries worldwide. These active Amazon marketplaces serve hundreds of millions of customers annually, offering a monumental product range and inventory. They allow consumers to buy just about anything, including clothing, beauty supplies, gourmet food, jewelry, books, movies, electronics, pet supplies, furniture, toys, garden supplies and household goods.</p> <p>In more recent years, Amazon gained prominence as a <a href="https://www.techtarget.com/searchitchannel/definition/cloud-service-provider-cloud-provider">cloud service provider</a>. The company launched Amazon Web Services (<a href="https://www.techtarget.com/searchaws/definition/Amazon-Web-Services">AWS</a>) in 2006 to provide cloud computing services like compute, storage, databases and networking to organizations. Through these offerings, along with emerging technologies like AI, <a href="https://www.techtarget.com/searchdatamanagement/definition/data-lake">data lakes</a>, analytics and IoT, Amazon aims to help companies and government agencies lower costs and enhance their agility and competitiveness.</p> <p>Apart from its main offerings -- e-commerce platform and 200+ cloud services -- Amazon is known for many of its other products, some of which are very popular. These include Twitch (a live streaming service), Kindle (e-book reader), Fire (tablets and TVs) and Alexa (smart speakers). The company is highly regarded for its ability to innovate and for being highly <a href="https://www.techtarget.com/searchcustomerexperience/feature/10-customer-service-best-practices-to-follow">customer-centric</a>.</p> </section> <section class="section main-article-chapter" data-menu-title="History and timeline of Amazon"> <h2 class="section-title"><i class="icon" data-icon="1"></i>History and timeline of Amazon</h2> <p>Amazon has come a long way since it was founded by Jeff Bezos in his garage in Bellevue, Wash., on July 5, 1994.</p> <p>The following is a brief history and timeline of events that led to Amazon's evolution from its humble beginnings to a multinational business empire.</p> <h3>The 1990s</h3> <p>Amazon officially opened for business as an <a href="https://www.techtarget.com/searchaws/feature/Amazons-impact-on-publishing-transforms-the-book-industry">online bookseller</a> on July 16, 1995, just one year after Bezos founded the company in his garage. For a few years, he shipped books to all 50 U.S. states and many other countries.</p> <p>Originally, Bezos incorporated the company's name as Cadabra, but later changed it to Amazon. He is said to have browsed a dictionary for a word beginning with A for the value of alphabetic placement. He ultimately decided on the name Amazon<i> </i>because it was exotic and different, and as a reference to his plan for the company's size to reflect the vastness of the Amazon River, one of the largest rivers in the world.</p> <p>Throughout 1995 and the first two quarters of 1996, Amazon was a loss-making company. However, in quarters three and four, <a href="https://www.techtarget.com/esg-global/blog/amazons-strategy-of-operational-excellence/">its</a> revenue rose and its losses fell. The company's losses continued to decrease over the next few years.</p> <p>In the second half of the 1990s, Amazon achieved several milestones. It launched an affiliate program to increase its reach in 1996 and opened its first remote distribution center in 1997. In 1998, Amazon made its first acquisition -- Internet Movie Database (IMDb) -- a website that provides information about movies and TV shows and is now used by millions of people worldwide.</p> <h3>The early 2000s</h3> <p>In October 2000, Amazon launched its first overseas operations in Japan. The Japanese Amazon marketplace remains accessible at <a target="_blank" href="http://www.amazon.co.jp/" rel="noopener">www.amazon.co.jp</a>.</p> <p>In the last quarter of 2001, Amazon turned a profit for the first time in its history. Then, 2003 became its first profitable year.</p> <p>Throughout the 2000s, Amazon expanded its original e-commerce bookstore to include hundreds of other products in many other categories, such as software, personal care, music goods, gourmet foods, sporting goods, and photography items like cameras. The expansion of its product range continued over the next few years, with the addition of newer categories like jewelry, baby products and vehicles.</p> <p>By the middle of the decade, Amazon was thinking beyond e-commerce. In 2005, it launched <a href="https://www.techtarget.com/searchapparchitecture/tip/Architecting-beyond-microservices-and-monoliths">Amazon Prime</a>, a membership-based service that offers free two-day shipping to Amazon customers in many countries, including the U.S., as well as streaming, shopping and reading benefits. The service has significantly contributed to Amazon's increasing revenue and profits since 2005.</p> <p>In 2006, Amazon made a successful foray into <a href="https://www.techtarget.com/searchcloudcomputing/tip/Explore-the-pros-and-cons-of-cloud-computing">cloud computing</a>, with the launch of AWS. Supported by an extensive global <a href="https://www.techtarget.com/searchcloudcomputing/definition/cloud-infrastructure">cloud infrastructure</a>, AWS offers a highly flexible and secure cloud computing environment for organizations in many industries, including automotive, energy, financial services, healthcare, manufacturing, education and government. That same year, Amazon also launched a video-on-demand service known at the time as Unbox.</p> <p>The company launched the Kindle e-reader in 2007. The same way that amazon.com changed the way people bought books, Kindle reshaped how they read them. This device helps users browse, buy and read e-books, magazines and newspapers from the Kindle Store.</p> <p>In the late 2000s, Amazon launched the Kindle app for <a href="https://www.techtarget.com/whatis/definition/mobile-device">mobile devices</a> (2009) and started Amazon Studios (2010) to develop original movies and TV shows. It also made numerous acquisitions like the audiobook retailer Audible.com (2008) and shoe retailer <a href="https://www.techtarget.com/whatis/feature/Best-cities-for-tech-workers-with-short-commutes">Zappos.com</a> (2009).</p> <h3>From the 2010s to present day</h3> <p>Amazon debuted its first tablet computer, the Kindle Fire, in 2011 and the Amazon Fire TV Stick, which is part of Amazon's extensive line of <a href="https://www.techtarget.com/whatis/definition/streaming-media">streaming media</a> devices, in 2014.<br><br>In 2011, the company created a Kindle lending library for Prime members. In 2012, it added a publishing arm with the acquisition of Avalon Books and opened its first <a href="https://www.techtarget.com/whatis/feature/Black-Friday-statistics">Black Friday</a>-deals store in 2013. It also started an online Amazon Art marketplace for fine arts in 2013, which has featured original works by famous artists such as Claude Monet and Norman Rockwell.<br><br>In 2014, Amazon launched the Echo smart device embedded with a voice-activated <a href="https://www.techtarget.com/searchcustomerexperience/definition/chatbot">chatbot</a> named Alexa. Alexa was rolled out to consumers in 2015 and was followed by the Alexa-equipped <a href="https://www.techtarget.com/searchaws/definition/Amazon-Echo">Echo</a> Dot in 2016. In 2025, Alexa powers millions of devices worldwide, as well as many Amazon services like Amazon Prime Video, Amazon Music, and Amazon e-commerce.</p> <p>Also in 2014, Jeff Bezos bought the <i>Washington Post</i> for $250 million. He continues to own the newspaper as of March 2025. However, several scandals have been associated with his ownership, resulting in staff exits and discontent. These problems have led to significant financial losses for the paper and have eroded its readership base.</p> <p>Amazon continued its acquisition and product launch spree throughout the decade. In 2017, it forayed into online groceries by <a href="https://www.techtarget.com/searchaws/feature/Amazon-grocery-strategy-could-shake-up-food-retail-industry">acquiring Whole Foods</a>. It also launched <a href="https://www.computerweekly.com/feature/Amazon-Go-is-now-the-right-time">Amazon Go</a>, a chain of cashierless grocery stores in 2018.</p> <p>The rise of in-home shopping during the COVID-19 pandemic of 2020 and 2021 made consumers rely on Amazon's e-commerce platform even more. With an estimated 310 million active users worldwide, it remains a popular choice for online shopping and its popularity is likely to endure in the coming years.</p> <p>In 2023, Amazon launched Amazon Bedrock, a fully managed service to build and scale generative AI (<a href="https://www.techtarget.com/searchenterpriseai/definition/generative-AI">GenAI</a>) applications. Then, in late 2024, Amazon announced that it would invest $4 billion in AI start-up Anthropic. The strategic collaboration between the two companies aims to enhance Amazon Bedrock's performance, security and privacy capabilities. In the long term, the partnership is expected to strengthen Amazon's GenAI capabilities, particularly in the cloud computing space.</p> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineimages/amazons_key_milestones-f.png"> <img data-src="https://www.techtarget.com/rms/onlineimages/amazons_key_milestones-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineimages/amazons_key_milestones-f_mobile.png 960w,https://www.techtarget.com/rms/onlineimages/amazons_key_milestones-f.png 1280w" alt="Timeline of Amazon's key milestones." height="427" width="559"> <figcaption> <i class="icon pictures" data-icon="z"></i>A look at Amazon's history and the evolution of its notable products and services. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="What is AWS?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What is AWS?</h2> <p>AWS is a comprehensive and evolving cloud computing platform. Amazon launched the AWS cloud to provide organizations with essential computing services through the internet. By doing so, they could avoid the burden of IT infrastructure management and easily access sophisticated technologies that were previously out of their reach.</p> <p>The first AWS offerings, Amazon Simple Storage Service (<a href="https://www.techtarget.com/searchaws/definition/Amazon-Simple-Storage-Service-Amazon-S3">S3</a>) and Amazon Elastic Compute Cloud (<a href="https://www.techtarget.com/searchaws/definition/Amazon-Elastic-Compute-Cloud-Amazon-EC2">EC2</a>) were launched in 2006 (a few months apart from each other) to provide online services for websites and client-side applications. Amazon S3 enables organizations to safely store data while maintaining privacy and control, while Amazon EC2 allows them to access massive amounts of computer processing power without setting up their own expensive data center. Amazon S3 and Amazon EC2 remain the backbone of the company's growing collection of web services. <br><br>AWS is one of Amazon's most lucrative businesses, generating segment sales of $28.8 billion in 2024 -- a 19% year-over-year increase from 2023. In 2025, the AWS ecosystem includes more than 200 cloud-based services used by millions of customers worldwide. Some of the most popular AWS services include:</p> <ul class="default-list"> <li><b>Amazon Elastic Block Store. </b><a href="https://www.techtarget.com/searchstorage/tip/Amazon-EBS-vs-EFS-An-elastic-comparison">EBS</a> is a block-storage service for Amazon EC2.</li> <li><b>Amazon CloudWatch. </b>This service monitors applications and gets insights into the operational health of systems.</li> <li><b>Amazon Relational Database Service. </b>RDS is a service to automate <a href="https://www.techtarget.com/searchdatamanagement/definition/relational-database">relational database</a> management tasks.</li> <li><b>AWS Lambda. </b>This is<b> </b>a <a href="https://www.techtarget.com/searchitoperations/definition/serverless-computing">serverless compute</a> service to run code in response to events without having to manage servers or <a href="https://www.techtarget.com/whatis/definition/cluster">clusters</a>.</li> <li><b>Amazon CloudFormation. </b>This tool provisions infrastructure as code (<a href="https://www.techtarget.com/searchitoperations/definition/Infrastructure-as-Code-IAC">IaC</a>) to speed up deployment and simplify management.</li> <li><b>Amazon Elastic Load Balancing. </b>ELB is<b> </b>a network service that automatically distributes incoming application traffic to improve network and application <a href="https://www.techtarget.com/searchdatacenter/definition/scalability">scalability</a>.</li> </ul> <p>The AWS ecosystem includes tens of thousands of partners that offer numerous products and services that integrate with AWS to support thousands of use cases. All partners in the Amazon Partner Network (<a href="https://www.techtarget.com/searchaws/definition/AWS-Partner-Network-APN">APN</a>) are listed on the <a target="_blank" href="https://partners.amazonaws.com/search/partners" rel="noopener">AWS Partner Portal</a>. Through this portal, any organization can find an AWS-approved provider to meet its cloud computing needs.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/-xanQ3aUWms?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.techtarget.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> </section> <section class="section main-article-chapter" data-menu-title="Notable Amazon products and services"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Notable Amazon products and services</h2> <p>Amazon offers an ever-expanding portfolio of services and products. Following is a list of its noteworthy offerings.</p> <h3>Retail</h3> <ul class="default-list"> <li><b>Amazon Marketplace.</b> Amazon's e-commerce platform enables third-party retailers to showcase and sell their products alongside Amazon items.</li> <li><b>Amazon Fresh.</b> Amazon's grocery pickup and delivery service is currently available in nearly two dozen U.S. cities and a few international locations. A grocery order can be placed through the Amazon Fresh website or the Amazon mobile app. Customers can either get their groceries delivered or visit the store for pickup.</li> <li><b>Amazon Vine.</b> Launched in 2007, Amazon Vine is an invitation-only program in which Amazon selects certain reviewers, sends them certain items free of charge, and invites them to share their experiences using those products. Vine helps manufacturers and publishers get reviews for their products to <a href="https://www.techtarget.com/whatis/feature/Ways-to-prevent-tech-buyers-remorse">help shoppers make informed purchases</a>.</li> <li><b>Woot.</b> Founded in 2004 and acquired by Amazon in 2010, Woot offers limited-time offers and special deals on many product categories like clothing, electronics and houseware. Many of these deals rotate daily, and some new deals are offered every 30 minutes.</li> <li><b>Zappos.</b> Amazon bought Zappos in 2009. This online retailer of shoes and clothing carries a wide range of brands, including Nike, Sperry, Adidas and UGG.</li> <li><b>Amazon Merch on Demand.</b> This <a href="https://www.techtarget.com/searchitoperations/definition/on-demand-computing">on-demand</a> printing service enables sellers to create and upload their designs for free. Amazon prints the merchandise -- T-shirts, accessories, etc. -- and delivers them to customers. The creators earn royalties on each sale.</li> <li><b>Amazon Handmade.</b> This platform enables artisans to sell handcrafted products to customers around the world.</li> </ul> <h3>Consumer technology</h3> <ul class="default-list"> <li><b>Amazon Kindle.</b> Kindle is Amazon's first e-reader. It enables users to browse, buy and read e-books, magazines and newspapers from the Kindle Store.</li> <li><b>Amazon Fire tablet.</b> Previously known as Kindle Fire, Amazon's popular Fire tablet competes with Apple's iPad.</li> <li><b>Amazon Fire TV.</b> This line of streaming media players and digital devices delivers streamed video content over the internet to a paired <a href="https://www.techtarget.com/whatis/definition/HDTV-high-definition-television">high-definition television</a>. These devices integrate with many streaming services including Amazon's own Prime Video, as well as third-party services like Netflix and Hulu.</li> <li><b>Amazon Fire TV Stick.</b> This compact, plugin streaming device connects to a TV's High-Definition Multimedia Interface, or <a href="https://www.techtarget.com/whatis/definition/HDMI">HDMI</a>, port to display streamed content from various services.</li> <li><b>Amazon Alexa.</b> This cloud-based, <a href="https://www.techtarget.com/searchcustomerexperience/definition/virtual-assistant-AI-assistant">AI-powered, voice-controlled personal assistant</a> is designed to answer queries, interact with users, and perform other tasks and commands. Amazon offers tools, application programming interfaces (<a href="https://www.techtarget.com/searchapparchitecture/definition/application-program-interface-API">APIs</a>) and reference solutions to help developers build applications for Alexa.</li> <li><b>Amazon Echo.</b> This is one of Amazon's <a href="https://www.techtarget.com/iotagenda/definition/smart-home-or-building">smart home</a> devices that comes equipped with a speaker. Alexa can listen to a user's voice commands and perform several functions in response, such as providing information about the weather, creating shopping lists and controlling other smart products, such as lights, switches and televisions.</li> <li><b>Amazon Echo Dot.</b> A smaller, puck-shaped version of the original Amazon Echo speaker, an Echo Dot speaker can be placed in any room and can answer questions, play music, and read news and other stories.</li> <li><b>Amazon Echo Show.</b> Part of the Amazon Echo line of speakers, Amazon Echo Show works similarly through Alexa. It also offers a 7-inch touchscreen display to play videos and music and conduct video calls with other Echo users.</li> <li><b>Amazon Astro.</b> This is Amazon's first home monitoring <a href="https://www.techtarget.com/whatis/definition/robotics">robot</a> that works with Alexa. It allows users to check on the rooms, people or things in their home. The robot can also help with various other household tasks, such as caring for the elderly through notifications and alerts, and following owners from room to room to play TV shows, music or podcasts.</li> </ul> <h3>Subscription services</h3> <ul class="default-list"> <li><b>Amazon Prime.</b> This <a href="https://www.techtarget.com/whatis/feature/Rise-of-the-subscription-economy-What-it-is-and-how-it-works">subscription service</a> provides members access to exclusive shopping and entertainment services, discounts and more. As an example, all Amazon Prime members enjoy free one-day or two-day shipping on qualifying orders.</li> <li><b>Amazon Prime Video.</b> This is Amazon's on-demand <a href="https://www.techtarget.com/searchunifiedcommunications/definition/streaming-video">video streaming</a> service that offers a selection of thousands of movies and TV shows. The service, included with an Amazon Prime membership, is available in over 200 countries and territories worldwide.</li> <li><b>Amazon Drive.</b> Previously known as Amazon Cloud Drive, Amazon Drive was a <a href="https://www.techtarget.com/searchstorage/definition/cloud-storage">cloud storage</a> app that offered limited free and secure online storage for photos, videos and files for Amazon customers. The service was discontinued as of Dec. 31, 2023, and as a result, customers no longer have access to their files on Amazon Drive. However, they can use the Amazon Photos service instead.</li> <li><b>Twitch Prime.</b> Twitch is a video streaming platform that offers a fun and social way to watch people play games, cook, create crafts, and so on. Amazon purchased Twitch in 2014.</li> <li><b>Amazon Prime Music.</b> This is Amazon's music and <a href="https://www.techtarget.com/searchunifiedcommunications/definition/podcasting">podcast</a> streaming service. It is free for Prime members.</li> </ul> <h3>Digital content</h3> <ul class="default-list"> <li><b>Amazon Pay.</b> An <a href="https://www.techtarget.com/searchdatacenter/definition/OLTP">online transaction processing</a> platform, Amazon Pay enables Amazon account holders to use their Amazon accounts to pay external online merchants, pay bills, book tickets, and buy foods, medicines, insurance, gift cards and gift vouchers.</li> <li><b>Kindle Store.</b> Kindle Store is Amazon's e-commerce store for buying e-books. Accessible from any Kindle device, it uses a recommendation engine to make personalized e-book recommendations to each buyer.</li> <li><b>Amazon Appstore for Android.</b> Amazon's app store for the <a href="https://www.techtarget.com/searchmobilecomputing/definition/Android-OS">Android operating system</a> enables users to download games and mobile apps to supported devices. However, on Aug. 20, 2025, Amazon will discontinue this service.</li> </ul> <h3>AWS</h3> <p>In addition to the many popular services highlighted previously, the growing AWS universe includes the following products:</p> <ul class="default-list"> <li><b>Amazon S3.</b> This is Amazon's scalable, cloud-based <a href="https://www.techtarget.com/searchstorage/definition/object-storage">object storage</a> service based on a pay-as-you-go model. It provides a scalable way to store any amount of data. In S3, files are known as <i>objects</i> and stored in containers called <i>buckets</i>.</li> <li><b>Amazon Simple Queue Service.</b> Amazon <a href="https://www.techtarget.com/searchaws/definition/Amazon-Simple-Queue-Service-SQS">SQS</a> is a fully managed message queuing service for <a href="https://www.techtarget.com/searchapparchitecture/definition/microservices">microservices</a>, distributed systems and serverless applications. It uses first-in-first-out (FIFO) queues to ensure that messages sent to systems are always published in the correct order.</li> <li><b>Amazon EC2.</b> Amazon EC2 is Amazon's web-based compute service for <a href="https://www.techtarget.com/searchdatacenter/definition/workload">workloads</a>. It allows users to access scalable infrastructure on demand. They can also run virtual servers or <a href="https://www.techtarget.com/whatis/definition/instance"><i>instances</i></a> that can be scaled up or down to optimize performance and cost.</li> <li><b>Amazon S3 Glacier storage classes. </b><a href="https://www.techtarget.com/searchaws/definition/Glacier-Amazon-Glacier">Amazon S3 Glacier</a> provides storage classes for data archiving. Different classes are available for different access patterns and storage duration.</li> <li><b>AWS Identity and Access Management.</b> AWS IAM enables <a href="https://www.techtarget.com/searchcloudcomputing/tip/An-introduction-to-AWS-IAM-best-practices">secure and controlled access</a> to AWS resources and services. It is particularly useful to apply fine-grained permissions, securely scale access across AWS accounts and establish preventive security guardrails on AWS.</li> <li><b>Amazon Redshift.</b> This cloud-based <a href="https://www.techtarget.com/searchdatamanagement/definition/data-warehouse">data warehouse</a> service enables users to query petabytes of both structured and semistructured data using standard SQL queries. Redshift analyzes data across data warehouses, operational databases and data lakes, and provides near real-time analytics to support organizational decision-making.</li> </ul> <h3>Amazon AI services</h3> <ul class="default-list"> <li><b>Amazon SageMaker.</b> A fully managed cloud <a href="https://www.techtarget.com/searchenterpriseai/definition/machine-learning-ML">machine learning</a> platform, <a href="https://www.techtarget.com/searchaws/definition/Amazon-SageMaker">Amazon SageMaker</a> enables developers and data scientists to build, train and deploy <a href="https://www.techtarget.com/searchenterpriseai/feature/How-to-build-a-machine-learning-model-in-7-steps">machine learning models</a>. It provides unified access to all data and allows users to use familiar AWS tools for model development, data processing and SQL analytics.</li> <li><b>Amazon Lex.</b> Amazon Lex is an AI chatbot builder. Users can use Lex to build conversational interfaces into any application. It is powered by the same technology as Alexa and uses advanced natural <a href="https://www.techtarget.com/searchenterpriseai/definition/language-modeling">language models</a> to build AI chatbots and voice bots in applications.</li> <li><b>Amazon Polly.</b> This fully managed service allows users to deploy high-quality, natural-sounding human voices in many different languages. It uses <a href="https://www.techtarget.com/searchenterpriseai/definition/deep-learning-deep-neural-network">deep learning</a> technology to convert any text to an audio stream, including text from articles, web pages and PDF documents.</li> <li><b>Amazon Rekognition.</b> This <a href="https://www.techtarget.com/searchenterpriseai/definition/facial-recognition">facial recognition</a> platform uses deep learning to process images and videos and extract information from them. It includes a facial verification feature to detect real users and spoofs. It can also compare faces, extract skewed and distorted text from images and detect unsafe or inappropriate content in images and videos.</li> <li><b>AWS DeepLens.</b> DeepLens was a deep learning-enabled video camera that developers used to build machine learning applications. Amazon ended support for DeepLens on Jan. 31, 2024, and deleted all DeepLens models, projects, and device information from the DeepLens service so users can no longer access it from the AWS console.</li> <li><b>Alexa Voice Service.</b> AVS-powered Alexa built-in devices provided developers with a set of APIs, software development kits, or <a href="https://www.techtarget.com/whatis/definition/software-developers-kit-SDK">SDKs</a>, and documentation to add Amazon Alexa's speech and other capabilities to their applications and devices. However, these tools are no longer available. To build devices that connect to Alexa, developers must now sign up for the Works with Alexa (WWA) certification program.</li> <li><b>Amazon Transcribe.</b> This automatic speech recognition (ASR) service automatically converts speech to text. Powered by a next-gen parameter speech foundation model, Transcribe allows developers to add speech-to-text capabilities to their applications.</li> <li><b>Amazon Translate.</b> <a href="https://www.techtarget.com/searchaws/definition/Amazon-Translate">Amazon Translate</a> can translate large amounts of text from one language to another. It uses machine learning to deliver high-quality translations of user-generated content, online conversations and more.</li> <li><b>Alexa Skills Kit.</b> ASK provides self-service APIs, Skill components, and other tools to build Skills (apps for Alexa).</li> </ul> <h3>Amazon privately owned brands</h3> <ul class="default-list"> <li><b>AmazonBasics.</b> This is Amazon's privately labeled, low-budget brand that mainly sells everyday consumer goods like kitchen, tech, office supplies and household products.</li> <li><b>Mama Bear.</b> This private label of Amazon sells baby wipes, diapers, baby food, vitamins, baby bottles, pacifiers and feeders.</li> <li><b>Presto!.</b> This brand started as a laundry detergent in 2016 but has since added many other household products to its product line. These include paper towels, toilet paper, disinfectants, cleaning liquids and surface cleaning wipes.</li> <li><b>Amazon Essentials.</b> The <a href="https://www.techtarget.com/searchaws/feature/Break-down-the-components-of-Amazons-retail-business">Amazon Essentials</a><b> </b>clothing line offers clothes, footwear and accessories for men, women, babies and kids.</li> <li><b>Happy Belly.</b> Introduced in 2016, this is Amazon's private label for food. Its product line includes snacks, teas, coffees, juices and canned fruits.</li> <li><b>Goodthreads.</b> This apparel line offers casual and professional clothes for both men and women.</li> </ul> <figure class="main-article-image full-col" data-img-fullsize="https://www.techtarget.com/rms/onlineImages/aws-amazons_business_web-f.png"> <img data-src="https://www.techtarget.com/rms/onlineImages/aws-amazons_business_web-f_mobile.png" class="lazy" data-srcset="https://www.techtarget.com/rms/onlineImages/aws-amazons_business_web-f_mobile.png 960w,https://www.techtarget.com/rms/onlineImages/aws-amazons_business_web-f.png 1280w" alt="Graphic highlighting Amazon's business web." height="319" width="560"> <figcaption> <i class="icon pictures" data-icon="z"></i>From online shopping to subscription services to publishing, Amazon offers products and services in a multitude of industries. </figcaption> <div class="main-article-image-enlarge"> <i class="icon" data-icon="w"></i> </div> </figure> </section> <section class="section main-article-chapter" data-menu-title="Notable Amazon subsidiaries and acquisitions"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Notable Amazon subsidiaries and acquisitions</h2> <p>From healthcare to entertainment, Amazon has acquired multiple companies in a variety of sectors over time.</p> <p>Following is a list of Amazon's notable acquisitions and subsidiary companies:</p> <ul class="default-list"> <li><b>IMDb. </b>The world's most popular database for movies, TV, celebrity, video games and streaming online content was acquired by Amazon in 1998.</li> <li><b>Audible. </b>Audible, a book and spoken audio content provider, was acquired by Amazon in 2008 for $300 million.</li> <li><b>Zappos. </b>Amazon acquired this online shoe and clothing retailer in an all-stock deal worth $1.2 billion in 2009.</li> <li><b>Twitch. </b>A social media and video game streaming platform, Twitch was purchased by Amazon for $970 million in 2014.</li> <li><b>Whole Foods. </b>Food, beverage and organic grocery store chain Whole Foods was acquired by Amazon for $13.7 billion in 2017.</li> <li><b>Ring. </b>Amazon took ownership of this home security and smart home company in 2018 for $1 billion.</li> <li><b>Zoox. </b>An <a href="https://www.techtarget.com/searchenterpriseai/definition/driverless-car">autonomous vehicles</a>, robotics and transportation company was acquired as a wholly owned subsidiary by Amazon for $1.2 billion in 2020.</li> <li><b>Metro-Goldwyn-Mayer. </b>Amazon acquired this film and TV studio for $8.5 billion in March 2022.</li> </ul> <h3>Amazon controversies and criticisms</h3> <p>Amazon has suffered a massive backlash over the years from multiple sources. The tech giant is also being held responsible for creating the <i>Amazon effect</i> -- the evolution and <a href="https://www.techtarget.com/searchaws/feature/Is-construction-Amazons-next-target-for-disruption">disruption</a> of the retail market due to the company's monopolistic behaviors.</p> <p>Following are a few concerns and allegations that Amazon has faced over time:</p> <ul class="default-list"> <li><b>Monopolistic and anticompetitive behavior. </b>Due to Amazon's size and economies of scale, it has <a href="https://www.techtarget.com/searchaws/feature/5-stories-on-the-Amazon-effect-across-industries-and-society">continuously outpriced local and small shopkeepers in many countries</a>. It is also accused of displacing an open market with a privately controlled one. This is leading to the slow death of the brick-and-mortar store model built.</li> <li><b>Unfair treatment of workers. </b>Amazon is frequently criticized for promoting <a target="_blank" href="https://www.forbes.com/sites/jackkelly/2021/10/25/a-hard-hitting-investigative-report-into-amazon-shows-that-workers-needs-were-neglected-in-favor-of-getting-goods-delivered-quickly/?sh=137f754d51f5" rel="noopener">unfair work conditions</a> in its warehouses, including treating workers as robots, providing low wages and creating unsafe work conditions.</li> <li><b>Huge carbon footprint.</b> Over the past two decades, Amazon has been accused by environmental activists of having a staggering <a href="https://www.techtarget.com/whatis/definition/carbon-footprint">carbon footprint</a>. Transport of any merchandise relies on oil, and since Amazon delivers almost anything to just about everywhere, it leaves a long-lasting carbon footprint.</li> <li><b>E-waste. </b>Amazon is criticized for contributing to the world's <a href="https://www.techtarget.com/sustainability/definition/e-waste">e-waste</a> crisis by destroying millions of unused or returned products. This also includes millions of electronics, such as phones, computers and TVs that are toxic to soil, water, air and can potentially harm wildlife, flora and delicate natural ecosystems.</li> <li><b>Counterfeit product listings.</b> Brands, shoppers and lawmakers have scrutinized Amazon because counterfeiters have been listing and selling fake products on Amazon. To crack down on counterfeit products, Amazon destroyed 2 million counterfeit products sent to its warehouses and blocked 10 billion fake listings in 2021. The company has also built tools to identify individual units of enrolled products to <a href="https://www.techtarget.com/searchcio/feature/Top-10-benefits-of-blockchain-technology-for-business">ensure that only authentic products</a> reach customers.</li> <li><b>Avoiding taxation. </b>Edging fast toward a monopoly status, Amazon has been criticized for often avoiding tax payments despite making huge profits. According to a report by the Institute on Taxation and Economic Policy, the company avoided around $5.2 billion in corporate federal income taxes in 2021. Many tax authorities have investigated Amazon for tax evasion. Some have let the company off the hook. For example, in December 2023, the EU's top court ruled that Amazon won't have to pay $273 million in back taxes.</li> </ul> </section> <section class="section main-article-chapter" data-menu-title="Amazon finances"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Amazon finances</h2> <p>In 30 years, Amazon has grown into a business behemoth with a presence in multiple industries and the revenue to match.</p> <p>As of the last quarter of 2024, AWS held the largest market share in the cloud infrastructure market. It is trailed by <a href="https://www.techtarget.com/searchcloudcomputing/definition/Windows-Azure">Microsoft Azure</a> and <a href="https://www.techtarget.com/searchcloudcomputing/definition/Google-Cloud-Platform">Google Cloud</a>. In addition, it reported revenue of $20 billion in this period, a substantial increase from the $10.6 billion earned in the last quarter of 2023.</p> <p>In all of 2024, the company's net sales increased to $638.0 billion -- an 11% increase over the 2023 figure of $574.8 billion.<b> </b>Net income also<b> </b>increased to $59.2 billion, compared with $30.4 billion in 2023.</p> <p>Here are some other notable statistics from a news release published by Amazon on Feb. 6, 2025:</p> <ul class="default-list"> <li><b>Net sales. </b>Increased 10% to $187.8 billion in the fourth quarter of 2024, compared with $170 billion in the fourth quarter of 2023.</li> <li><b>Operating income.</b> Increased to $21.2 billion in Q4 2024, compared with $13.2 billion in Q4 2023.</li> <li><b>Net income. </b>In Q4 2024 increased to $20.0 billion, almost twice the number of $10.6 billion in Q4 2023.</li> <li><b>Operating income. </b>Increased from $36.9 billion in 2023 to $68.6 billion in 2024.</li> <li><b>Operating cash flow. </b>Increased by 36% from $84.9 billion in 2023 to $115.9 billion in 2024.</li> </ul> <p><i>Besides being recognized as a company with business interests in e-commerce, cloud computing and AI services, Amazon also offers an extensive list of </i><a href="https://www.techtarget.com/searchaws/feature/Learn-the-scope-of-Amazon-subscription-services"><i>subscription services</i></a><i>. Learn about these services and the perks they offer.</i></p> </section> Amazon (Amazon.com) is the world's largest online retailer and one of the largest providers of cloud services. https://cdn.ttgtmedia.com/visuals/digdeeper/3.jpg https://www.techtarget.com/whatis/definition/Amazon Wed, 23 Apr 2025 09:00:00 GMT 60 webmaster@techtarget.com