Overview of IBM Db2 AI for z/OS
IBM Db2 AI for z/OS (“Db2ZAI”) helps customers improve the operational performance and maintain the health of their Db2 for z/OS system by offering features like SQL optimization, System assessment, (augmented by Performance insights) and Distributed connection control ("DCC").
SQL optimization empowers the optimizer in your Db2 for z/OS engine to determine the best-performing query access paths, based on your workload characteristics. The access path for an SQL statement dictates how Db2 for z/OS accesses the data that the query specifies. It determines the indexes and tables that are accessed, the access methods that are used, and the order in which objects are accessed.
As part of the training process, Db2ZAI collects data from the Db2 for z/OS optimizer and the query execution history, which are derived from workloads in your unique operating environment. Once trained, the model is ready to be deployed into production, providing insights to the optimizer's access path selection. These insights are in addition to what the optimizer uses today in the selection of the best query path. The information is unique to your environment, and currently unknown to the traditional query optimizer.
Internal IBM benchmarks have shown performance benefits in terms of reduction in CPU and elapsed time associated with these types of queries.
In the area of Db2 system performance, the System assessment feature of Db2ZAI (augmented by Performance insights) helps you tackle the problems of interpreting and acting upon the enormous amounts of performance data produced by Db2 for z/OS. Database administrators and performance analysts often do not have the time or skills needed to fully use the metrics that Db2 provides to tune the system or identify early symptoms of performance issues. Db2ZAI leverages machine learning technology to discover non-optimal performance events, and then provides a set of recommended actions to improve the situation. The Performance insights feature integrates new application-level performance data and provides graphical reports with drill-down capabilities for deeper understanding.
Distributed applications that rely on the rich data that exists in Db2 for z/OS can scale to support large enterprise operations. The large volume of connection requests from these distributed applications can impact other distributed applications and, in the worst-case scenario, impact the availability of the z/OS system. DCC helps mitigate the risks associated with distributed connection requests coming into the Db2 for z/OS system by allowing users to set warnings and exceptions related to these incoming connections so that distributed applications are fenced versus the whole business running on the Db2 system getting impacted. The DCC feature leverages AI to learn the distributed connection behavior to control the flood of connections and threads using Db2 profile function. It recommends Db2 profile controls to prevent a remote application from impacting other remote applications and shows users the detailed thread and connection information so they can better manage connections. Its recommended profiles that create a penalty box when a remote application starts to create an unusually large number of connections or threads until the cause is determined. This helps ensure Db2 for z/OS systems continue to perform reliably and efficiently under the most demanding hybrid cloud environments.
Large portions of Db2ZAI processes are eligible to run on IBM® Z Integrated Information Processor (zIIP), resulting in minimal impact to general processor CPU.
