With new Fabric features, Microsoft aims at AI development
Microsoft on Tuesday introduced new features for Fabric to enable customers to develop AI-powered applications.
First unveiled in May 2023 and made generally available six months later, Microsoft Fabric is an AI-fueled data management and analytics suite that packages previously disparate platforms Data Factory, Azure Synapse Analytics and Power BI together in a single environment.
In combination, the suite is intended to enable seven data management and analytics workloads, including real-time BI, data engineering, data warehousing and data science.
New features designed to fuel AI application development include Fabric Databases to provide users with intuitive and autonomous databases for feeding applications with relevant data — the first of which is SQL Database in Microsoft Fabric — and AI Functions in Fabric to analyze text.
In addition, Microsoft unveiled OneLake Catalog, a data catalog for Fabric that enables data governance and exploration, and new features for databases SQL Server and Redis, among other new features.
The tech giant revealed all the new capabilities — most of which are in preview — during Microsoft Ignite, a user conference held both in person in Chicago and online.
Together — along with other AI capabilities unveiled at Ignite, such as AI Azure AI Foundry Service, AI Model Gallery and Agent Framework — the new data management and analytics capabilities have the potential to bring Microsoft’s AI capabilities more in line with those of competitors such as AWS and Google, according to Doug Henschen, an analyst at Constellation Research.
He noted that Microsoft was an early developer of AI-powered assistants in 2023. But the tech giant took a long time to make those assistants generally available, and as AI has evolved over the past year to become more agentic, Microsoft has lagged.
“The company took a long, long time to deliver Copilots and has been caught rather flat-footed on the push toward agentic applications,” Henschen said. “We’ll see if the [new capabilities] being introduced at Ignite can quickly gain traction. For now, AWS, Google Cloud and apps competitors like Salesforce have drawn a lot of attention on the path forward on AI.”
While perhaps Microsoft is no longer the fastest to provide AI capabilities and foster AI development, the tech giant’s new data management and analytics features are nevertheless important for its customers. Combined, they advance Microsoft Fabric as an infrastructure for developing AI, according to Kevin Petrie, an analyst at BARC U.S.
“This is a pretty comprehensive set of enhancements,” he said. “Microsoft continues to strengthen its data fabric as a foundation for generative AI and [traditional] AI, from the infrastructure to data and application layers.”
When Microsoft hosted Build, a user conference for developers, in May, new features for Fabric included support for open-source table format Apache Iceberg, an AI assistant in Power BI and improved real-time analytics capabilities.
New capabilities
Enterprise interest in developing AI applications, including generative AI, has surged in the two years since OpenAI’s November 2022 launch of ChatGPT, which marked a significant advancement in large language model technology.
Two of generative AI’s main benefits are its potential to enable widespread use of analytics to inform decisions with true natural language processing and its ability to take on repetitive tasks previously performed by human beings. Essentially, it holds the promise of making people both smarter and more efficient.
As a result, many enterprises have begun developing AI-powered applications or expressed interest in doing so.
In response, with data the foundation of any AI model or application, many data management and analytics vendors have developed generative AI-powered tools and created environments for customers to develop AI tools of their own.
Microsoft was one of the first, forming a partnership with OpenAI in early 2023 that included a $10 billion investment in the AI vendor. Since then, AI has been a focus for Microsoft, according to Frank Shaw, the tech giant’s chief communications officer.
“We’re innovating at every level of the Copilot and AI stack, spanning developer tools, AI platforms and data infrastructure, all so that customers can build their own AI applications and infuse AI in every aspect of their business,” he said in a virtual press briefing on Nov. 15 to introduce the 90 new features unveiled during Ignite.
New capabilities in Fabric, including Fabric Databases and AI Functions in Fabric, both of which are in public preview, further Microsoft’s development of an ecosystem for its users to quickly and easily build AI applications.
Fabric Databases is designed to simplify AI application creation by providing developers with autonomous databases — databases that use AI and machine learning to automate tasks — that intuitively make relevant data available as they build new applications.
SQL Database in Microsoft Fabric, the first Fabric database to be introduced, is a database engine that enables users to develop a unified data platform that applies AI across both operational data and analytical data. Operational data includes such information as individual transaction records while analytical data tracks patterns in those individual records over time.
Features of SQL Database in Microsoft Fabric include data that is automatically replicated to OneLake so it is available to Fabric analytical engines, native vector search combined with embedding models in Azure AI to enable retrieval-augmented generation, support for continuous integration and continuous delivery (CI/CD), and Copilot in Fabric so users can interact with data using natural language.
AI Functions in Fabric enables users to analyze text, providing summarization, translation and sentiment analysis, among other capabilities.
Microsoft Fabric initially focused on analytical data, Henschen noted. The addition of support for developing transactional data applications on the same platform is therefore significant, he said
In addition, the combination of operational and analytical data is extended further by automatically replicating data to OneLake and a new feature called Open Mirroring in Microsoft Fabric that enables users to automatically replicate data from any application or other data source as the data changes, Henschen continued.
“The addition of a [transactional] option and support for change data capture replication rounds out Fabric as a platform for data-driven applications as well as data-driven insights,” he said.
However, Microsoft’s new data replication and capabilities to capture data changes are not unique, Henschen added, noting that AWS offers something similar.
“These options are very much like the … ties AWS has introduced between database services such as Amazon Aurora and Amazon Redshift,” he said.
Other new Microsoft data management and analytics capabilities include the following:
- OneLake Catalog, a data catalog that enables users to govern and explore their entire Microsoft Fabric ecosystem of structured and unstructured data, data lakehouses and warehouses, machine learning models, notebooks, reports, dashboards and other data and AI assets. OneLake Catalog’s exploration features are generally available while its governance will be in public preview in early 2025.
- SQL Server 2025, an updated version of the SQL Server database now in private preview that includes features aimed at enabling AI development such as vector storage and a Copilot that aids SQL development by offering suggestions, completing code and providing best practices.
- Azure Managed Redis, a service now in public preview that will automatically deliver the latest developments from the open source Redis data store, including capabilities that cache data more efficiently to boost performance and lower costs.
While each feature provides new functionality, the OneLake Catalog provides particularly valuable capabilities, according to Henschen.
Microsoft provides enterprise-wide data cataloging and metadata management capabilities through Purview. OneLake Catalog focuses those capabilities on Microsoft Fabric’s data lake to provide organizations with trusted data on which to develop AI applications and information about their data and AI assets.
“OneLake Catalog … brings all the data exploration and governance goodness of cataloging and metadata management to OneLake,” Henschen said. “Here’s a very focused, easy to implement and manage set of capabilities just for OneLake.”
Petrie likewise called out the OneLake Catalog. Before the launch of the new data catalog, Microsoft only provided Fabric users with a data hub. The data catalog gives Fabric users more advanced governance and data discovery capabilities.
“The new OneLake Catalog marks a good step forward for Microsoft users that need to document, find and govern elements across their analytics projects — data, models, platforms,” Petrie said.
Beyond simply aiding Microsoft users, the new data management and analytics capabilities aid the tech giant in its ongoing competition with fellow hyper-scalers, he continued.
Unlike platforms that were built for the cloud from the start, Microsoft served an on-premises audience of users before developing cloud-native capabilities. Now, many enterprises prefer hybrid deployments that enable them to keep their most sensitive information on premises and other data in the cloud.
With its SQL focus and other features of Fabric, Microsoft is perhaps better positioned to serve the needs of those customers than some of its competitors.
“Microsoft is capitalizing on one of its primary advantages over Databricks, Snowflake and the cloud hyper-scalers — namely, longstanding support for hybrid environments,” Petrie said. “Many SQL environments include data and applications that remain on premises. Unlike its major competitors, Microsoft Fabric helps manage the on-prem elements.”
Next steps
With most of the capabilities unveiled during Inspire now in preview, features such as Fabric Databases, including SQL database in Microsoft Fabric, essentially make up Microsoft’s data management and analytics roadmap.
At their core — and at the core of many of the nearly 100 new features the tech giant unveiled on Tuesday — is enabling Microsoft Fabric users to develop AI applications that make users more productive, according to Shaw.
“It’s our vision to provide every employee with a Copilot and a team of agents working with them and on their behalf,” he said.
Henschen, meanwhile, said that beyond introducing SQL Database in Microsoft Fabric, he expects to see more than just one Fabric Database.
Like AWS, Microsoft already has an array of database offerings on Azure. But Fabric Databases promises to be different than other Microsoft databases by specifically integrating them with Fabric.
“Fabric Databases sounds like a promise to deliver more than just one generic SQL database, so I won’t be surprised to see more database options in the future, akin to the many options AWS offers through its Relational Database Service offerings,” Henschen said.
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.
#Fabric #features #Microsoft #aims #development