E-Handbook: Understand how the edge and IoT intertwine Article 2 of 4

iconimage - Fotolia

How to build an effective edge IoT architecture

Explore edge IoT architectural components, what each part does and how they all connect to each other, including sensors, actuators, devices and gateways.

An IoT network has been compared to a tech version of a living entity because its sensors see, hear and feel the environment around it. Just as humans react to the information we get from our senses -- such as sweating when we are overheated -- IoT networks use hardware and software components to communicate and work together seamlessly.

As the number of connected sensors and mobile devices grows, IoT architectures must evolve. Today's IT architectures simply can't compete with the volume of data generated by IoT devices and maintain the processing power required to transmit and process that data. The only way to keep up is to integrate edge computing with IoT architecture.

Edge IoT architecture moves network connections from centralized resources to the resources, communications and traditional processing more closely located to the devices producing the data in order to sustain the data volume and velocity. The move to the edge focuses on prioritization and optimization of data transmission for the entire network.

This type of architecture distributes intelligence across the IoT network to boost performance, bandwidth, efficiency, reliability and security.

What is edge IoT architecture?

Edge computing is a distributed IT architecture in which data is processed as close to the originating source as possible. In an IoT architecture, that means processing data closer to the IoT hardware and not sending it to the cloud or an on-premises data center for processing.

Example of an IoT gateway system

An edge IoT architecture encompasses everything from the endpoint of the system -- including devices, sensors, actuators and appliances -- to the IoT gateway. The gateway is the communications hub for the IoT network. It performs critical functions for the network, such as sensor data aggregation, translation between sensor protocols, and sensor data processing before the gateway forwards the data to the cloud or on-premises network.

Given the explosion of IoT device deployments in the last few years, having the right edge IoT architecture is a critical part of a highly available and high-functioning IoT network.

The importance of a proper edge IoT architecture

IoT devices generate massive amounts of data and require processes to execute in milliseconds. According to Gartner, approximately 10% of enterprise-generated data today is created and processed outside of a traditional data center or the cloud. By 2025, it predicts this figure to reach 75%. IDC has noticed similar trends and believes the global datasphere will grow to 175 zettabytes by 2025.

Global datasphere

In an IoT-connected office building environment, thousands of sensors monitor temperature, light, noise, air quality and security systems. Actuators in key systems respond to changes detected by the sensors to turn lights off and on, unlock security doors or adjust the HVAC system based on outside weather. Each IoT device and actuator may use different protocols -- such as Wi-Fi, Bluetooth, serial ports, Message Queuing Telemetry Transport (MQTT) or Zigbee -- to connect to the network, and may have different management and security models.

A well-architected edge IoT network can handle the influx of data from all devices as well as a portion of the data processing. However, some functions, like unlocking a security door, can't wait for the time it would take for the data to make the round trip from the IoT device to a cloud-based application or data center. Other edge devices -- such as a low-power controller that is used infrequently -- are not able to analyze the data they collect because they lack the computing power to do so, while still others would generate so much data that it would not be valuable in its raw form.

The benefits of a proper edge IoT architecture

A well-implemented edge IoT architecture reduces a system's overall latency, increasing performance for high-use applications and devices. It also reduces data and communication bottlenecks at architectural integration points between the IoT network and other services or parts of the network, such as the cloud, so it can function at the speed employees and customers demand today.

A proper edge IoT architecture can also help organizations:

  • rapidly and efficiently deploy IoT projects;
  • reduce security risks of the IoT network overall;
  • scale IoT projects horizontally and vertically within a tech stack; and
  • add newer processing technologies such as machine learning and AI to an IoT network.

That's not to say that deploying an edge IoT architecture will work for every organization. Organizations that implement the technology without an adequate security or IT infrastructure strategy will struggle to keep up and potentially introduce more risk than they would solve.

IT professionals and technology leaders must learn to juggle competing business objectives with the disruptive technologies deployed today. According to Ross Winser, a senior director analyst at Gartner, traditional tools and processes are quickly reaching their limits. By involving IT pros and leaders early in IoT architecture planning discussions, he wrote in a blog post, organizations can avoid the cascade effect of unforeseen service gaps that could cause serious headaches in the future.

What do you need in your edge IoT architecture?

Designing and implementing an edge IoT architecture depends on numerous factors:

  • the IoT devices, including their age, integration technology and communication protocols;
  • how much intelligence needs to be deployed at the edge;
  • the physical location of the devices; and
  • the systems affected by edge IoT devices.

At a basic level, edge IoT architecture -- not including the data center, cloud and enterprise layer -- will consist of two layers: the edge device and middle server.

Edge device layer

The device tier, or edge, is the outermost layer of IoT architecture. It consists of various hardware components, including sensors, devices and actuators.

Sensors monitor the real-world status of connected products and machines. While most people think of them as only physical objects, anything that gathers and transmits data can be considered a sensor -- for example, light, temperature and moisture sensors, GPS receivers and vehicle onboard diagnostic sensors.

When talking about IoT, people typically think of hardware devices with sensors to activate and control them and transmit IoT data. Examples include microcontrollers embedded inside smart vehicles and household appliances, and single-board computers such as Raspberry Pi devices embedded in robotics or industrial devices.

Actuators control or react to something in the IoT system. They can be either hardware or software that affects the physical state of a product or the environment. APIs trigger actuators based on data gathered from the rest of the IoT architecture, such as smart lights, remote-controlled access points, motors, robotics and software commands.

Middle server layer

Also known as the control tier, this layer of the IoT architecture contains the IoT gateway and other servers that facilitate communication, offload processing functions and drive actions at the edge. It's a place to locally preprocess the data gathered from the edge before sending it to the data center or cloud.

Data preprocessing alleviates bandwidth constraints and optimizes the amount of data forwarded to the rest of the system. Minimizing the amount of data transmitted affects network transmission costs, especially over cellular networks, resulting in significant savings. Most gateways enable bidirectional communication and prioritize communications based on defined rules and guidelines. For example, gateways can hold all southbound data transmission while a critical device update package comes the other way.

Further, it allows critical business rules and filters to be applied to the data in real time. Data is processed, transmitted or stored for later use based on these filters, enabling companies to manage data more efficiently and develop more useful insights.

A founding pillar

Edge computing is still developing, but it's fast becoming one of the pillars of every organization's IT infrastructure. The number of devices and sensors being deployed on the edge device layer increases every day. IT professionals must know how that influences their overall architecture.

Device choice and placement become more vital given the dependence systems and organizations have on that data to operate. Organizations find new and innovative ways to use IoT, and only architectures with a solid foundation will be able to scale and support edge IoT technology appropriately.

Dig Deeper on Enterprise internet of things

Data Center
Data Management