Microsoft has introduced a new database service called Azure HorizonDB, designed to handle the demanding needs of modern AI and cloud-native applications. This latest addition to Microsoft’s PostgreSQL family aims to deliver high performance and easy scalability, making it ideal for workloads that require quick responses and the ability to grow seamlessly.

Many enterprises rely on data storage solutions that can work across different environments and scales. While options like MySQL and proprietary systems like SQL Server exist, they often fall short when managing large, distributed databases or when operating across multiple cloud providers. PostgreSQL, an open-source database that’s been around for nearly 30 years, has become a popular choice because of its flexibility and support for a wide range of workloads. Microsoft has supported PostgreSQL on Azure since 2017, building from its acquisition of Citus Data in 2019, which enhanced its scaling capabilities.

What Makes Azure HorizonDB Different

Azure HorizonDB is built to be a high-performance, scalable database that addresses the limitations of previous PostgreSQL implementations. It’s designed for workloads that need to handle large amounts of data quickly and efficiently. Microsoft claims HorizonDB can increase throughput three times over standard open-source PostgreSQL when handling transactional tasks. It can scale up to 3,072 cores and handle 128TB of data, with sub-millisecond latency for commits.

The architecture of HorizonDB separates compute resources from storage. This means if you need more processing power or additional read replicas, HorizonDB can provision those resources independently. This flexibility is crucial for AI applications, large Kubernetes deployments, and other cloud-native workloads. Data is replicated across multiple regions and availability zones, ensuring high availability and durability, with automated backups and maintenance that minimize disruptions.

Built for AI and Modern Workloads

Microsoft has added features to HorizonDB that support AI applications. For instance, it offers fast vector search capabilities based on DiskANN technology, which is essential for retrieval-augmented generation (RAG) tasks and model tuning. These advanced search functions can significantly boost performance compared to standard vector searches in PostgreSQL, especially when working with large indexes stored across memory and disk.

The database also supports bringing AI models directly into SQL queries. Microsoft Foundry, a tool integrated with HorizonDB, allows users to run AI operations within their database queries. Managed models are available through the Azure Portal, or users can bring their own. This setup makes it easy to generate embeddings, perform semantic searches, or summarize data without leaving the database environment. Thota, a Microsoft executive, explains that users can simply keep their existing SQL structures and invoke semantic operators to enhance their data insights.

Security is a priority with HorizonDB. It integrates with Azure’s enterprise security tools, offering encryption, private endpoints, and support for Azure Entra ID. Azure Defender for Cloud provides added protection for sensitive data, making HorizonDB suitable for enterprise use.

Tools for Developers and Migration Support

Microsoft also announced a new version of its Visual Studio Code extension for managing PostgreSQL databases. This tool works with HorizonDB as well as other PostgreSQL instances, whether hosted on-premises or in the cloud. It provides visual tools to explore database schemas, monitor performance metrics, and optimize operations—all within a familiar development environment.

One of the key features is support for AI-assisted database development. Developers can use natural language queries to design databases or refine existing queries, leveraging GitHub Copilot integration. This makes working with complex data structures more accessible, especially for those new to database management.

Another major focus is helping organizations migrate from other relational databases, especially Oracle. The tools in the VS Code extension can assist in converting Oracle schemas to PostgreSQL, using AI to handle transformations based on best practices. Users can validate the migration in a sandbox environment before deploying, reducing risks and ensuring compatibility. While some manual adjustments may still be needed, this process aims to simplify and accelerate database migration projects.

In summary, Microsoft’s Azure HorizonDB represents a major step forward for scalable, AI-ready PostgreSQL solutions in the cloud. Its architecture, performance features, and developer tools are designed to meet the needs of modern enterprises looking to leverage data at scale and integrate AI capabilities seamlessly.

Inspired by

• https://www.infoworld.com/article/4093191/azure-horizondb-microsoft-goes-big-with-postgresql.html

Sources

• gmpg.org
• microsoft.com
• microsoft.com
• microsoft.com
• microsoft.com