MongoDB on Thursday unveiled a host of new features for its database, highlighted by new data migration and generative AI capabilities.
The new features -- some of which are generally available while others are in preview -- were revealed during MongoDB.local NYC, an in-person user event for application developers in New York City, where the vendor is based. In addition, MongoDB launched updated versions of numerous existing tools.
As a group, the new and updated capabilities represent mostly incremental progress for MongoDB, according to Stephen Catanzano, an analyst at TechTarget's Enterprise Strategy Group.
However, the addition of generative AI capabilities through a partnership with Google Cloud and a tool called MongoDB Atlas Vector Search stand apart, he noted.
"The generative AI announcement would be considered innovative," Catanzano said. "MongoDB's audience is very developer centric. And with Google Cloud and the tools they are enabling [with generative AI] to attract developers, this should get the attention of developers trying to quickly integrate generative AI. The rest is mostly moving the ball forward."
Relational databases struggle to discover relationship between data points, which is a necessity as organizations ingest data from an ever increasing number of sources. In addition, because they were first invented in 1970, the databases were not built for the scale many organizations now require as their data volume continues to grow.
As a result, alternatives such as graph databases that specialize in discovering relationships between data points are becoming more popular. MongoDB, meanwhile, is a document-based database and is designed to work with large sets of distributed data.
Last year, MongoDB launched version 6.0 of its database, which included a new queryable encryption feature that enables users to search encrypted data and an updated version of its time series capabilities to increase the speed with which it can read data.
MongoDB and generative AI
Generative AI has the potential to be a transformative technology for data scientists, data engineers and other data experts.
Historically, their work has required the use of complex code.
Whether constructing data pipelines, developing applications or training models, they've had to go through the painstaking process of coding every command. In addition, when glitches and other anomalies have arisen, they've had to manually search for the source of the error and write new code to fix it.
The large language model capabilities of generative AI tools reduce the need to painstakingly write code. Their extensive vocabularies allow for true natural language processing, converting written words to code that software systems understand.
Stephen CatanzanoAnalyst, Enterprise Strategy Group
In addition, their machine learning capabilities enable the automation of certain previously manual tasks.
The result is increased productivity for data workers no longer forced to spend much of their time on repetitive, time-consuming tasks. They're free to do deep analysis, which subsequently boosts the productivity of their organization.
As a result, many data management and analytics vendors have unveiled capabilities incorporating generative AI since OpenAI's launch of ChatGPT in November 2022, which represented a significant advancement in generative AI and LLM capabilities.
Already, in June, vendors including Dremio, MicroStrategy and Monte Carlo have unveiled tools incorporating generative AI.
Now MongoDB is adding generative AI capabilities to its database through an extension of its partnership with Google Cloud that will enable developers to use Google generative AI and large language model capabilities as they build applications with MongoDB Atlas.
In addition, MongoDB revealed Atlas Vector Search in public preview to enable organizations to quickly and easily build applications that include generative AI and LLM capabilities aimed at improving productivity with ease of use.
Vectors are numerical representations of unstructured data, such as text, images and videos, that essentially bring structure to the unstructured data points. That subsequently enables previously unstructured data to be combined with structured data and used to inform data applications.
As a result, MongoDB's foray into generative AI is significant, according to Rachel Stephens, an analyst at RedMonk.
"Vector search is a hot new category in this era of generative AI," she said. "We've seen new special-purpose vector database companies raise significant VC money in recent months [while] existing database companies like MongoDB are also announcing vector search capabilities in an effort to integrate vector search into their tools."
Workload migration and more features
As organizations start to realize the limitations of relational databases, many are trying to migrate their data workloads to more modern data repositories, according to Andrew Davidson, MongoDB's senior vice president of products, who spoke during a virtual press conference on June 20.
MongoDB wants to be the destination for those organization trying to leave relational databases behind, he continued.
Toward that end, the vendor launched in general availability MongoDB Relational Migrator to simplify the migration of application and data transformation workloads from relational databases to MongoDB's document-based database.
"The Relational Migrator is an important announcement to help accelerate data migrations," Catanzano said.
In addition to new generative AI and data migration capabilities, MongoDB's new and updated tools include the following:
- MongoDB Atlas Stream Processing available in private preview later in 2023 to improve streaming data processing from sources such as IoT devices, end-user browsing behavior and inventory feeds.
- A preview of MongoDB 7.0, which is scheduled for release later this summer. It includes new options for deploying MongoDB on AWS, expanded support for the programming language Kotlin, and simplified data processing and analytics capabilities using Python.
- New MongoDB Atlas Search Nodes to help customers scale search workloads in applications independent of their database and enable users to isolate workloads, optimize their resources and improve performance at scale.
- Improved scaling and flexibility using MongoDB Time Series collections aimed at simplifying enterprise-scale time series workloads that can grow quickly as devices such as IoT sensors and users' browsers send data into a database for processing.
- Support for Microsoft Azure in MongoDB Atlas Online Archive and Atlas Data Federation, both of which already support AWS.
Beyond new and updated tools, MongoDB unveiled Atlas for Industries, a program that will develop industry-specific versions of the vendor's developer platform.
Data management specialists, including Databricks and Snowflake, have each made developing industry-specific versions of their platforms a priority over the past two years, releasing versions targeted at organizations in such industry verticals as manufacturing and healthcare.
MongoDB's first industry-specific version of Atlas is designed for financial institutions and is now available.
Stephens noted that despite adding support for search, time series and geospatial data in recent years, MongoDB is still looked at largely as a document database.
However, the new features now available plus those in preview show the versatility of the vendor's database.
"While the company added support for other data types in recent years, the market perception still tends to consider MongoDB primarily a document database," she said. "These announcements show concerted effort by the company to continue to expand into multi-modal workloads."
A common thread
All of the new and updated tools were designed with the common goal of attracting more workloads from existing customers and enticing new users to migrate their workloads to the vendor's database, according Davidson.
During MongoDB's most recent fiscal quarter, the vendor added about 2,300 new customers to bring its total number of customers over 43,000, president and CEO Dev Ittycheria reported during MongoDB's quarterly earnings call on June 2.
"We're focused on winning more workloads, bringing them onto the platform and making our customers successful with each of them," Davidson said.
As MongoDB looks forward, one way the vendor might continue to attract new customers and expand use by existing customers is by addressing real-time data management, according to Catanzano.
He noted that the vendor's database does not excel at managing data in real time. As more organizations recognize the benefits of real-time data, MongoDB would be wise to improve its real-time data management capabilities.
"Real-time data management is what a lot of companies are focused on," Catanzano said. "MongoDB is not a leader in this space. More focus on that message and capability will become important."
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.