SHARE San Jose 2017 - EWCP Focus on Capacity Management and Performance

It’s difficult to see everything with so many sessions to choose from at SHARE.  Making it even more difficult this year was EWCP having multiple sessions at the same time.  My favorite session for the over 25 years has been the EWCP project opening w/ the WSC zOS Hot Topics update.  This is a must see for anyone who cares about performance or getting more capacity out of current or future hardware.  It’s also a great session to find out what is new that you want to see in other EWCP sessions.  Below is a summary of some sessions that I attended that will be of value to you, if you are not familiar with the topic.  Many of these sessions have important updates, which are listed in the summary.

20274 EWCP Open / WSC Hot topics - Just to name a few things covered; zIIP Honor Priority by Service class, z13 Hiperdispatch chip cache efficiency improvements.  But this session also includes important APARs for issues or performance.

20273 - WLM Update for Mobile Pricing, Capping  - Mobile usage pricing now much easier as WLM classification rule creates SMF 70 Mobile MSUs for SCRT.  Two new Absolute Caps also discussed.  Sysplex routing now aware of LPAR time to cap to avoid sending more work.

20143 &20269 z/OS SMT Measurements & Deciding Whether to Enable - Two excellent sessions on how to understand zIIP constraints and how SMT might be right for you to get some more capacity instantly.  If you are overflowing zIIP to GCP it’s costing MLC $.

20127 Reduce z13 MLC with Processor Cache Optimization - Repeat of a Best of SHARE session.  A must see for anyone considering or running on a z13 processor.  Excellent session to ensure your workload gets the rated MSU capacity of the box you are considering.  Huge swings in effective MSUs possible based on your workloads RNI. Also see session 20321 if interested in RNI.

20314   Capping, Capping, and Capping - Comparison of Capacity Controls a more in-depth look at all the PR/SM and WLM capping controls.  Excellent session from IBM with updates for the new Absolute capping options.

20619   Improving XCF performance to keep up with CPC improvements - Many customers ignore XCF performance or manage to ancient guidelines.  If your XCFAS CPU is high, the problem is much bigger than you think. This session talks about simple ways to check performance, new guidelines and how to easily improve XCF performance and reduce its CPU / MSU impact. 

20386 WLM Velocity Goals in the Year 2017 - Excellent session by Peter Enrico a frequent EWCP speaker.  He discusses why our old velocity goals are obsolete due to advances in hardware and why response goals should be reviewed as well.

20320 Coupling Technology What's the Right Stuff for Me? You can often save GCP MSUs w/ CF tuning to reduce SYNC service times.  But there is something new to totally avoid making SYNC CF calls, DB2 GBPs duplexing supports ASYNC w/ z13.

20321 Relatively New LSPR and the Latest z Systems Processors - Must see for those who are considering z13 migration and haven't heard of RNI.  RNI effects # MSUs consumed for same instructions / workloads.

20337   How to Measure that New IBM z13 - After learning RNI concepts in Session 20127 above, find out how to measure how your current z13 is performing.  Also recommend past SHARE EWCP sessions on using zPCR for RNI.

20606 Wells Fargo's Experience with DB2 z/OS Buffer Pool Simulation   - While not a EWCP session, it has interesting z/OS Performance information.  How to use real memory effectively to reduce z/OS I/O and CF Synchronous requests to DB2 Group buffer pools. Both save GCP MSUs and make workloads run faster while consuming less billable MSUs.  After reading look at EWCP CF session listed above 20320 for maximum DB2 GBP MSU impact reduction.

1 Comment
1 Like

Nice article

April 11, 2017 04:04 PM by Meral Temel

Nice article Don, Thank you! . 

Recent Stories
SHARE in Pittsburgh August - MVSP Sessions / Speakers

SHARE in Sacramento - MVS Performance Project

SHARE Providence Aug - EWCP and z14