Boosting Database Performance in Oracle 21c Database with Persistent Memory.

The Latest Persistent Memory such as Intel Optane Persistent Memory (PMEM) combines near-DRAM performance with the data persistence of storage. It can be configured in memory mode or application direct mode. When the memory mode is used, PMEMs can be considered as volatile and can be used as main memory and DRAM is treated as a write-back cache. PMEMs can also be configured in application direct mode to store database files and redo logs. This presentation will discuss the use cases of PMEM:1) In memory mode use PMEMs for Oracle In-Memory store;2) in Application direct mode to store the database files, namely persistent Memory Database in Oracle 21c. We will discuss the significant performance benefits that can be achieved from either of these two modes. We will discuss the Persistent Memory Database feature that includes directly mapped buffer cache and Persistent Memory Filestore (PMEM Filestore) in Oracle 21c. We will also discuss how PMEMs are used in Oracle Exadata X8M.

In-database Machine Learning with Oracle Database and/or Oracle Autonomous Database

Many Machine Learning tasks need to access a lot of data set, which in many cases are stored in a database such as Oracle Database. It makes a more scalable solution to do the machine learning task in the database, which is called in-database machine learning. Oracle Autonomous Database comes with a library of Oracle machine learning algorithms and a set of building tools such as SQL notebooks for machine learning. This allows Data scientists to run Machine Learning projects in Oracle Database without moving data. This session will examine Oracle Machine Learning as part of the Oracle database as well as Oracle Autonomous Database. We will discuss the process of machine learning: analyze and prepare data set; build and evaluate and apply machine learning model. We also will discuss the Oracle machine learning features in Oracle 21c such as AutoML for In-Database Machine Learning and newly added in-database machine learning algorithms for anomaly detection, regression, and deep learning analysis.