Your overall consumption and its relationship to your billed MSU's can be understood by reviewing your cost10 (R4H) and cost14 (Billed MSU) in z/Trim. These two KPI's look at your system utilization a little differently focusing on your R4H (cost10) and the relationship of your R4H and WLM Soft Capping Support. This soft capping this is defined using DC/GC (IBM's Defined Capacity or Group Capacity) options.
If your IBM MLC is based solely upon your R4H, your billed MSU will match your R4H. are using based upon your R4H (Rolling Four-Hour Average).
When soft capping is used, the system behavior is modified. If the R4H exceeds the DC/GC limit, WLM will throttle the LPARs where this condition exists and there will be short periods when the LPARs will be denied CPU resources until the R4H drops below the defined DC/GC level.
The good news is that the billed MSU's can be maintained at a lower level to reduce operating costs. The bad news is that your application performance will be adversely affected by the effects of PR/SM capping and your application SLA's may be in jeopardy.
The first cost10 screenshot shows the cost10 R4H for 09/03/2020. The R4H is shown throughout the 24-hour period. The R4H peak is 1018 MSU's.
The cost10 KPI has the capability to present many different views of resource consumption and the R4H. You can setup up an instance of this KPI to present an analysis based upon LPAR, CPC, Sysplex or Entity. An entity represents a grouping of logical partitions (LPARs) that is typically done to map the LPARs to meaningful business units or cost centers. In our demo system, the LPARs are grouped together according to our demo system's lines of business. Our demo system is based upon a bank. The entities are Commercial Banking, Personal Banking and Financial Banking.
The second cost10 KPI screenshot shows the same day's R4H but is showing the analysis by entity.
This analysis provides information to size the consumption based upon the business workloads that are running on the assigned LPARs. This provides the mainframe consumption across all lines of business in one concise KPI.
The second screenshot shows the cost14 Billed MSU for the same day (09/03/2020). The R4H is shown throughout the 24-hour period as well as the Billed MSU
The Billed MSU is 'only' 977 MSU's. WLM Soft Capping Support is being used. The gaps between the R4H and the Billed MSU's show periods during the day when application performance is adversely affected by PR/SM capping. Performance issues and potentially missed SLA's may be a result of this condition. However, the MSU cost is lower.
The cost14 KPI also offers different analysis options to display the billed MSU's. The billed MSU's can be presented by CPC or by LPAR/Group Capacity. The second cost14 KPI screen shot below provides the analysis of the Billed MSU by LPAR/Group Capacity.
------------------------------
Mark Perillo
Rocket Software
------------------------------