putilov_denis - stock.adobe.com
Oracle launched a major update to its MySQL cloud service, known as HeatWave.
The tech giant said the update is generally available now.
Oracle first introduced HeatWave in December 2020; the platform was originally known as the MySQL Analytics Engine. HeatWave provides MySQL running as a service on Oracle Cloud Infrastructure (OCI) with the ability to run online analytics processing workloads (OLAP) alongside online transaction processing (OLTP).
With the update, Oracle introduced a MySQL Autopilot capability that aims to optimize query acceleration and database provisioning. HeatWave users will now also get the new MySQL scale-out data management capability that accelerates data loading.
Machine learning for MySQL Autopilot on HeatWave
A core part of the MySQL HeatWave update is machine learning to help enable the Autopilot functionality -- a positive development for Oracle users, according to Futurum Research analyst Ron Westfall.
"I view Oracle's organization-wide commitment to developing machine learning capabilities across its portfolio as paying dividends in delivering MySQL HeatWave innovations," Westfall said.
Westfall noted that the machine learning capabilities of MySQL Autopilot can sample data to provide recommendations on how the system can be optimized. He added that MySQL Autopilot brings auto provisioning, auto error recovery and auto scheduling, which are needed to meet today's system-wide automation and mixed OLTP/OLAP workload optimization requirements.
MySQL Autopilot to boost query acceleration on HeatWave
The Autopilot collects many statistics that the system uses to train the machine learning models, explained Nipun Agarwal, vice president of MySQL HeatWave advanced development at Oracle.
For example, MySQL HeatWave collects statistics on how long it takes for different types of queries to execute and how data was accessed.
Ron WestfallAnalyst, Futurum Research
With that information collected over time, the MySQL Autopilot system can make informed recommendations for optimization to accelerate queries. Oracle has developed new methods for data sampling and statistics extrapolation to build the machine learning models that underpin MySQL Autopilot, Agarwal said.
Oracle is already using machine learning capabilities with the Oracle Autonomous Database platform that also runs in Oracle's cloud. While teams can collaborate across the Oracle Database and MySQL on ideas and techniques, the MySQL Autopilot is specific to HeatWave, Agarwal noted.
MySQL Autopilot runs on HeatWave, but open source will still benefit
The core MySQL database is an open source technology, though the HeatWave service is proprietary and only runs on Oracle Cloud Infrastructure.
Agarwal explained that MySQL running in OCI is always running on the latest stable version of the open source project. HeatWave, with its analytics and now MySQL Autopilot, are additional capabilities that complement the core database.
While MySQL Autopilot only runs in Oracle's cloud, Agarwal noted that the open source project has received many improvements that help enable the new capabilities.
For instance, he noted that the SQL Optimizer in the MySQL open source project has been improved to help enable some of the faster query capabilities that are now part of MySQL HeatWave. The open source project now also has been improved to accelerate the ingestion speed at which the MySQL database ingests data.
"Any changes which are cloud-specific -- those obviously don't make sense to go in the open source project," Agarwal said. "But we are doing a lot of other enhancements."