HeatWave Release Notes
This topic introduces the different features of MySQL HeatWave.
MySQL HeatWave is ideal for data warehousing, real-time analytics, mixed workloads, transactional processing, and machine learning workloads when the main data source is the DB System without needing ETL.
When you submit a query to the DB System from a MySQL client or application, a query optimizer automatically determines whether the query should be offloaded to MySQL HeatWave Cluster for accelerated processing. This decision is based on two factors: first, whether all operators and functions referenced in the query are supported by MySQL HeatWave Cluster, and second, whether the estimated processing time with MySQL HeatWave Cluster is less than that with the MySQL instance. If both conditions are met, the query is sent to MySQL HeatWave Cluster for accelerated processing. Once the query is processed, the results are returned to the MySQL instance and the client or application that issued the query.
To learn more about MySQL, see MySQL 9.4 Reference Manual.
Real-time Analytics Without ETL
Analytics queries always access the latest data, as updates from transactions in MySQL automatically replicate in real time to MySQL HeatWave. A MySQL HeatWave Cluster can accommodate up to 512 nodes. The number of nodes required depends on the size of the data and the amount of compression achieved when loading data into MySQL HeatWave Cluster. Data loaded into MySQL HeatWave is automatically stored in Object Storage, allowing for quick reloading when MySQL HeatWave Cluster resumes after a pause or recovers from a cluster or node failure.
Transactional Processing
MySQL HeatWave is built on MySQL Enterprise Edition and provides the same robust transactional capabilities as a fully-managed cloud-based database service. Additionally, MySQL HeatWave supports JavaScript-stored programs, enabling you to express complex procedural logic within the database and seamlessly access MySQL datasets without incurring additional ETL costs.
Learn more about accelerating query processing with MySQL HeatWave .
MySQL HeatWave Lakehouse enables you to query structured and semi-structured data (such as Avro, CSV, JSON, and Parquet) and unstructured data (such as documents, text files, and presentations) stored in Object Storage.. The query processing occurs within the MySQL HeatWave Cluster, enabling you to utilize MySQL HeatWave for both MySQL and non-MySQL workloads. The data in Object Storage is not copied into the MySQL InnoDB storage engine, thus avoiding duplication of data.
The service reads data from Object Storage, transforms it into the memory-optimized MySQL HeatWave format, and stores it in MySQL HeatWave Storage Layer within Object Storage. This data is then loaded into the memory of the MySQL HeatWave Cluster, making it available for queries.
Learn more about loading data from Object Storage into MySQL HeatWave Cluster with Lakehouse and using Lakehouse with MySQL HeatWave AutoML.
MySQL HeatWave AutoML is a fully-managed, highly-scalable, and cost-efficient machine learning (ML) solution for data stored in both DB System and Object Storage. It offers a simple SQL interface for training and utilizing predictive machine learning models, making it accessible to both novice and experienced ML practitioners. No specialized knowledge, tools, or algorithms are required.
With MySQL HeatWave AutoML, you can train a model with a single call to a
stored routine. Similarly, generating predictions only requires
a single CALL or
SELECT statement, which can be
easily integrated into your applications.
One of the key advantages of MySQL HeatWave AutoML is that your data and models remain within the secure MySQL HeatWave environment, saving you time and effort. MySQL HeatWave distributes ML computations across its nodes to leverage scalability and take full advantage of its massively parallel processing capabilities.
Learn more about training and using machine learning models with MySQL HeatWave AutoML.
MySQL HeatWave GenAI offers integrated and automated generative AI featuring in-database large language models (LLMs), in-database embedding generation, in-database vector store, scale-out vector processing, and the ability to have contextual conversations in natural language. This enables you to leverage generative AI without needing AI expertise, transferring data, or incurring additional costs.
Additionally, MySQL HeatWave GenAI features MySQL HeatWave Chat, a chatbot that enhances the generative AI and vector search capabilities. This enables you to ask multiple follow-up questions about a specific topic within a single session. MySQL HeatWave Chat can also access information from documents stored in the vector store.
Learn more about performing AI-Powered search and content generation with MySQL HeatWave GenAI.