MySQL HeatWave is the only fully managed MySQL database service that combines transactions, analytics, machine learning, and GenAI services, without ETL. It also includes the capabilities to query data in object store and provides the best price performance in the industry. MySQL HeatWave is available on Oracle OCI, Amazon AWS, and Microsoft Azure.
This year marks an important milestone: 30 years of MySQL. Over the past three decades, MySQL has become one of the world’s most popular database, trusted by innovative organizations across every industry. From powering websites and applications to handling enterprise-scale workloads, MySQL has been at the core of countless data-driven solutions.
Introducing MySQL AI: A New Era of Intelligence
In a transformative move, we’re excited to announce MySQL AI—a suite of generative AI, machine learning and vector processing capabilities of MySQL HeatWave into MySQL Enterprise Edition. This leap enables developers to build rich AI/ML applications with MySQL in the development environment of their choice and deploy the applications on-premises or cloud.
Whether you’re building chatbots, enhancing customer experiences, streamlining internal operations, detecting financial fraud, or powering predictive maintenance, MySQL AI makes advanced AI accessible using data that is already in MySQL or enterprise data in file systems.
Migration from existing MySQL application to MySQL HeatWave
We are seeing tremendous growth in MySQL HeatWave, in large part driven by organizations migrating MySQL applications from on-premise or other cloud to MySQL HeatWave. This is driven by the following:
- Faster performance
- Lower cost of ownership
- Ability to run analytics, machine learning, and AI on MySQL data in a single service
By eliminating the need for separate analytics engines or vector databases, MySQL HeatWave allows teams to do more—with less complexity.
Announcements from Oracle AI World
At Oracle AI World this week, we announced several new features predominately in four categories:
- Enterprise features
- Developer-centric enhancements
- Features for advanced analytics and AI workloads
- Facilitating more applications that use data in object store
A new MySQL HeatWave Migration Assistant
Customers can now migrate existing on-premise MySQL applications to MySQL HeatWave in one simple step! MySQL HeatWave Migration Assistant handles schema, data, users, and replication setup with zero downtime. Users stay in control with progress updates, completion estimates, and detailed logs for easy troubleshooting. MySQL HeatWave migration tool is available in preview and will be publicly available soon.
MySQL Studio – A new graphical interface with a query editor, notebooks, visualizations of results
MySQL Studio is a new graphical interface that provides an intuitive, integrated environment for database management and application development. It includes a SQL worksheet, a chat interface for querying documents stored in the vector store and data in MySQL, and an interactive notebook for developing machine learning and generative AI applications. The notebooks are compatible with Jupyter, allowing developers to import existing notebooks for use with MySQL AI and MySQL HeatWave or share and collaborate on ML and GenAI projects.
NL2SQL, MCP, REST service – Faster and easier application development
Interact with Your Data Using Natural Language (NL2SQL)
NL2SQL in MySQL simplifies data access by letting developers ask questions in natural language, which are automatically translated into accurate SQL using large language models enriched with database schema context. Available in both MySQL AI and HeatWave, NL2SQL leverages in-database LLMs and MySQL enhancements that refine and augment LLM output for higher accuracy and performance. For business users, NL2SQL provides intuitive, self-service access to data—empowering them to make data-driven decisions without SQL expertise.
Build Smarter AI Apps with MySQL Model Context Protocol (MCP)
We have introduced MySQL Model Context Protocol (MCP) server and enhanced our Python SDK to enable developers to build AI-powered MySQL applications using popular open-source frameworks like LangChain. MCP standardizes how AI models interact with tools, data, and development environments by defining a unified approach to “tool-use” and “context” across clients (such as AI models or agents) and hosts (like IDEs, apps, or servers). With the MCP server and tools integrated into MySQL HeatWave, developers can streamline the creation of advanced agentic AI applications that harness MySQL HeatWave’s powerful capabilities.
Eliminate Middleware to Simplify Application Stack with MySQL HeatWave REST Service
We have introduced MySQL REST Service, which is native to MySQL and integrated with HeatWave. It lets you create RESTful API endpoints directly from your MySQL HeatWave instances—no backend code, middleware, or third-party frameworks required—reducing development and operational complexity. The service auto-generates REST endpoints mapped to your schemas and tables, enabling faster delivery of progressive web apps, AI-driven solutions, and dashboards. It’s secure and scalable, leveraging MySQL role-based access control, flexible authentication options, and a high-performance HTTP server.
Reduce Application Complexity with WebAssembly, and JSON Duality support in MySQL
MySQL HeatWave provides native support of JavaScript and now we have introduced WebAssembly (WASM) that run inside the database. This enables developers to create new applications with rich semantics and data processing capabilities. Developers can write stored functions and procedures using JavaScript libraries or modules compiled from C, C++, or Rust into WASM, supplementing the built-in SQL functions for more powerful operations.
JSON Duality View which was originally introduced in Oracle Database is now available in MySQL. It provides an innovative approach to managing document-centric applications without sacrificing the strengths of relational databases. It enables hybrid workloads combining SQL and document processing, smoother interoperability, and simplified development workflows, while still benefiting from MySQL’s robust features like referential integrity and strong transactional consistency.
- Link to blog on JSON Duality views in MySQL
- Link to blog on using JSON duality views in MySQL
- Demo video on creating and querying JSON Duality view in MySQL
Build Intelligent AI/ML Applications on-premise with MySQL AI
MySQL AI brings built-in generative AI, machine learning, vector creation/processing capabilities directly into the MySQL database, allowing developers to build intelligent applications using familiar tools—on any platform, including on-premises. With features like automated machine learning, built-in large language models (LLMs), and a vector store, MySQL AI eliminates the need for complex integrations or separate AI services. A novel aspect of MySQL AI is that it enables MySQL to perform semantics search on documents in file system.
Customers can either deploy applications built with MySQL AI on-premise running on existing hardware, or they can choose to run these applications with MySQL HeatWave for lower cost, better performance and richer functionality (Analytics, query data in object store, Autopilot). Migrating from MySQL AI to MySQL HeatWave is seamless, since it uses the same MySQL engine and supports existing protocols, tools, and frameworks—like MySQL Studio and the Model Context Protocol (MCP).
Improved HA, FSDR, BYOK, Autopilot Indexing – Robust and Secure OLTP
High Availability with 6x Better Failover Time
Modern enterprise applications demand a reliable and secure MySQL data platform capable of handling high-volume transactions with minimal downtime. MySQL HeatWave offers robust high availability (HA) features, and we’ve recently improved both planned and unplanned failover times by 6X. This enhancement ensures your applications remain resilient and responsive, reducing the risk of disruption and keeping critical services online.
Automated Disaster Recovery with Full Stack Disaster Recovery (FSDR)
Disaster recovery (DR) for OLTP workloads is crucial to ensure business continuity during system failures, data corruption, or natural disasters. With Oracle Full Stack Disaster Recovery (FSDR), organizations can automate end-to-end DR processes to minimize downtime and data loss. MySQL HeatWave integrates seamlessly with compute, storage, OKE clusters, and networking within full stack DR plans, simplifying setup and management. This orchestration of application stacks without manual intervention improves operational efficiency and gives you confidence that your critical applications can quickly recover from disruptions.
Improve operation efficiency with Autopilot Indexing and Notifications
Tuning database performance can be complex and time-consuming, but Autopilot Indexing takes the guesswork out of optimizing your OLTP workloads. It automatically analyzes query patterns and recommends the indexes (create/drop) for optimal workload performance while balancing storage overhead. What makes this feature even more powerful is the ability to automatically predict and notify users for the recommendation.
Customer-Managed Encryption Key for Enhanced Data Security
MySQL HeatWave encrypts all data both at rest and in transit by default. While Oracle manages encryption keys to ensure robust protection, organizations with stricter regulatory or internal requirements can choose to bring and control their own encryption keys. This Customer-Managed Encryption Key (CMEK) option provides full control over key management and data protection, giving businesses greater assurance and flexibility to meet their compliance and security policies.
New features for analytics, generative AI and querying object store
Build Real-Time Dashboards with Auto-Refresh Materialized Views and Temp Tables
MySQL HeatWave enables organizations to build advanced analytic applications by querying data in files in object store with record performance in the industry. We now introduced Auto-Refresh feature which propagates changes in files on object store to HeatWave automatically, thereby enabling real-time analytics.
Due to the compelling performance and cost of MySQL HeatWave, some customers migrate their workloads from non-MySQL and analytics databases. To provide a seamless migration and facilitate more warehouse use cases, we have introduced support for auto-refresh materialized views in HeatWave. Materialized views store precomputed query results directly in HeatWave memory and automatically update whenever the underlying tables change, ensuring your data is always fresh and accurate. This approach is perfect for complex analytics and dashboards, boosting system read scalability by reducing repetitive computations and resource consumption. Complementing this, we also introduced temp table support in HeatWave, which store data entirely in memory and provides SQL access to data. This results in significantly improved query performance and lower latency for applications with multi-step data transformations.
Support for Delta Lake in MySQL HeatWave
With MySQL HeatWave, you can query open file format such as Parquet, Avro, ND-JSON, and CSV. We now support Delta Lake files, a widely adopted open-source table format that brings ACID transactions, schema enforcement, and versioning to files in data lake. This combines the industry’s leading price-performance of HeatWave for PetaByte-scale lakehouses, and the transactional capabilities of delta lake format.
Accelerate Semantic Search with Automatic Vector Indexes in HeatWave
MySQL HeatWave supports semantic search by very efficient in-memory scan. To accelerate scenarios where data size is very large and some of the predicates are highly selective, we have introduced support for vector indexes. This provides up to 100x faster performance over traditional full scans. The vector index complements the existing efficient in-memory scan-based vector search, allowing HeatWave to intelligently choose the best method based on query context. Vector indexes in HeatWave are built automatically, requiring no manual setup or tuning which makes it easy to leverage HeatWave for GenAI workloads.
Unlock Precise and Relevant Search with HeatWave’s Hybrid Approach
MySQL HeatWave has added support for hybrid search which combines semantic and keyword-based approaches, making it possible to retrieve relevant information whether the user’s query is vague and conceptual or highly precise and fact-based. This is useful for business applications dealing with diverse datasets, from product catalogs with specific names/keywords to regulatory filings and customer interactions with detailed descriptions. This dual approach is crucial for successful Retrieval Augmented Generation workflows, where the combination of semantic relevance and exact matching ensures both accuracy and completeness in answers, and reduces hallucinations.
Summary
MySQL is one of the most popular databases in the world with tens of millions of applications and hundreds of millions of deployments worldwide. MySQL HeatWave is a fully managed database service from the MySQL team which is available in multiple clouds. Thousands of customers have migrated their MySQL applications to MySQL HeatWave for better performance, lower cost and enhanced OLTP, real-time Analytics, AL/ML and Lakehouse capabilities.
MySQL AI is an extension to MySQL EE and provides a rich set of GenAI, in-database LLMs, vector, and AutoML capabilities to MySQL which enables organizations to develop richer and more intelligent applications with MySQL.
In addition to MySQL AI, this week at Oracle AI World, we released a number of features in MySQL HeatWave targeting enterprise workloads, developers, AI/ML and Lakehouse applications. This we believe will provide customers even more compelling benefits of migrating MySQL workloads and developing new applications with MySQL HeatWave. MySQL HeatWave is now FedRAMP certified and is available in US Government and Department of Defense regions.