The annual North American IDUG conference was held in Anaheim, California and I always like going through the download of presentations to understand all the presentations I missed. Users, DB2 consultants and IBMers always provide great DB2 performance technical insights into how to get the most out of your environment.
This year the conference DB2 LUW tracks provided a number of great presentations highlighting the latest and greatest for getting the best DB2 performance tuning for LUW. Going through the conference presentations, one that sticks out is one on DB2 pureScale given by IBM’s Matt Huras, and Anixter’s Bernie O’Connor and Jason Spencer Within this presentation the combination of IBM architecture details and user Beta program experiences of the DB2 pureScale platform provides great information.
The 2-part presentation highlights the architecture of DB2 pureScale as well as its layout, configuration and monitoring. The DB2 performance and scalability provided by the multiple DB2 members attached to a pair of Cluster Caching Facilities (CF) provide the global locking and other facilities. The DB2 pureScale configuration looks just like its big brother’s z/OS DB2 Data Sharing configuration with its z/OS coupling facilities. DB2 pureScale also provides tremendous linear scalability for great DB2 performance.
In the second part of the presentation the user experiences show testing of the software member fail over, a software primary CF fail over and then a hardware member and CF fail over tests. These examples show how fast (4 seconds), DB2 pureScale LUW recovers and how DB2 performance is uninterrupted through these scenarios. DB2 performance within pure Scale just keeps rolling along.
Take the time to go through all the IDUG North American conference presentations and pick up all the latest DB2 performance tuning hints and tips, like this one on DB2 pureScale.
Leave a Reply