Microsoft Integrates Vector Search into Azure Cosmos DB for NoSQL

Apps & Games / Desktop / Windows / Microsoft Integrates Vector Search into Azure Cosmos DB for NoSQL
10 Jul 2024

Microsoft has unveiled a series of updates to its database offerings at the Build 2024 conference, with a notable enhancement being the integration of vector search into Azure Cosmos DB for NoSQL. This non-relational database service is part of the broader Azure Cosmos DB ecosystem, a distributed database that offers various consistency models, allowing users to balance performance and latency in their applications.

Expanding Capabilities in NoSQL

At its core, NoSQL databases eliminate the constraints of traditional SQL databases, such as data types and consistency, to achieve greater speed, flexibility, and scalability. Azure Cosmos DB supports a variety of data models, including APIs for MongoDB and Apache Cassandra, and also offers a PostgreSQL variant.

Last year, Microsoft introduced vector search to Cosmos DB, building on the Cosmos DB for MongoDB vCore service. The vector search in Cosmos DB for NoSQL is powered by DiskANN, a suite of scalable approximate nearest neighbor search algorithms that support real-time updates.

PostgreSQL and AI Integration

Additionally, Microsoft has made the Azure Database for PostgreSQL extension for Azure AI generally available. This extension enables developers who prefer PostgreSQL to directly integrate their data with Azure AI, simplifying the process of leveraging large language models (LLMs) and creating rich PostgreSQL generative AI experiences.

Moreover, the company is previewing an embeddings generation feature within its Azure Database for PostgreSQL offering. This new feature can generate embeddings directly within the database, enhancing the AI capabilities available to developers.

Enhancements to Copilot

Microsoft is also expanding the capabilities of Copilot within its databases. New features include the ability to provide summaries of technical documentation in response to user queries within Azure Database for MySQL. In March, the company announced a private preview of additional Copilot functionalities, further cementing its commitment to integrating AI-driven tools across its database services.

Update: 10 Jul 2024