Oracle’s new database uses machine learning to automate administration

Oracle’s new database

Oracle wants to make it easier for customers to power their businesses with its database. The company unveils a new 18c version of its marquee database software that uses machine learning. The new version protect customer data and automate the management of their information.

The software use machine learning on top of a wealth of log data to help optimize the database for frequent use patterns through caching, indexing, and other techniques. In addition, it will ensure that customers don’t have someone using stolen credentials to access business data. The database software will also expand or shrink the quantity of compute and storage it’s using automatically.

Administrators set policies for the database software to follow, and it will then manage itself following those rules.

“If you eliminate all human labor, you eliminate human error,” Oracle cofounder and CTO Larry Ellison said.

All that automation supposes to help free up human database administrators to work on other tasks, like planning and security which provide more value to its customers.

18c version

Oracle plans to launch 18c in December of this year for data warehouses, and in June of next year for online transaction processing workloads. Customers able to run the software in their private data centers on infrastructure they own, or have Oracle truck in its managed infrastructure through the company’s Cloud at Customer services. It will also available through the Oracle Cloud.

While, 18c available in all of those locations. Only the Cloud at Customer and public cloud versions of the software take advantage of the autonomous database features.

Existing Oracle customers can also benefit from a new set of discounts. Users can apply their existing database licenses to its platform-as-a-service offerings, like Autonomous Database Cloud.

Oracle could guarantee customers of the new software less than 30 minutes of planned or unplanned downtime per year. In addition, the performance tuning will consume less compute and storage than a human-managed database.