Mainframe CPC Specialty Engine Focus

This section contains the following topics:

RMFZAC - Daily zAAP Engline Use and Demand by CPC and SYSID

RMFZAS - Daily zAAP Engline Use and Demand by Service Class

RMFZIC - Daily zIIP Engline Use and Demand by CPC and SYSID

RMFZIS - Daily zIIP Engline Use and Demand by Service Class

RMFZAC – Daily zAAP Engine Use and Demand by CPC and SYSID

Overview

The RMFZAC query charts the hourly actual zAAP engine use, and zAAP engine eligible use for both the Central Processing Complex (CPC) and the z/OS LPARs using the shared engine pools of a CPC. Only PR/SM managed z/OS LPARs are charted. Non-z/OS LPARs and z/OS LPARs managed by VM are excluded from analysis because the data containing zAAP engine demand is not available.

The example shows the CPC level chart (top) for an IBM z/10 2097-712 CPC, and one of the individual LPAR charts for an LPAR defined to that CPC.

The CPC level chart shows total zAAP engine demand, with the actual use and eligible demand stacked. The horizontal reference line in the CPC level chart shows the number of physical zAAP engines in the CPC shared engine pool.

The LPAR level chart shows zAAP engine use and zAAP engine demand as separate vertical bars for each hour of the day. The horizontal reference line in the LPAR level chart shows the number of logical zAAP engines assigned to the LPAR. One LPAR level chart is produced for each LPAR that executed zAAP eligible work.

Uses

This query is used to determine if the zAAP engines in a CPC are sufficient to handle the zAAP eligible workload running on z/OS LPARs. Even if no zAAP engines are present in a CPC, the charts show z/OS LPAR zAAP eligible work that executed on CP engines. zAAP engines are less expensive that CP engines. Reducing CP engines in a CPC can reduce software licensing fees. In general, if significant zAAP eligible work is executing on CP engines, it is cost effective to add zAAP engines to the CPC.

Note: If a CPC has one or more zIIP engines and no zAAP engines, and has specified ZAAPZIIP=YES in SYS1.PARMLIB, zIIP engines process all specialty engine work—including work that would normally run on zAAP engines. In this environment, all SMF and RMF metrics treat zAAP and zIIP eligible workloads as zIIP eligible. No zAAP actual or eligible demand appears.

Data Extracts

Daily CPC zAAP Eng Use and Demand

Left Y-axis

ZAPENGA – zAAP Engine Actual Usage

ZAPENGE – zAAP Engine Eligible Usage

This extract shows the total z/OS LPAR zAAP usage at the CPC level. The vertical bars show total zAAP actual and zAAP eligible CPU time for all PR/SM managed z/OS LPARs at the CPC level. The CPU times are divided by 3600 to convert from seconds to hours. Because one engine can be dispatched a maximum of one hour, per hour, this conversion allows the CPU demand to be represented as physical engine dispatch hour equivalents. For example, two zAAP engines could, at full utilization, provide two hours of CPU time in one hour.

Right Y-axis

PRSMAZSP - Avg zAAP Shareable Processors

Average number of physical zAAP engines online in the shared zAAP engine pool.

X-axis

HOUR – Hour of Day

Daily z/OS System zAAP Eng Use and Demand

Left Y-axis

ZAPENGA – zAAP Engine Actual Usage

ZAPENGE – zAAP Engine Eligible Usage

This extract shows the z/OS LPAR zAAP usage for the CPCs charted in the CPC level data extract. The vertical bars show total zAAP actual and zAAP eligible CPU time for all PR/SM managed z/OS LPARs at the LPAR level. The CPU times are divided by 3600 to convert from seconds to hours. Because one engine can be dispatched a maximum of one hour, per hour, this conversion allows the CPU demand to be represented as physical engine dispatch hour equivalents. For example, two zAAP engines could, at full utilization, provide two hours of CPU time in one hour.

Right Y-axis

CPUAVOZP - Avg Number of Online zAAP Processors

Average number of logical zAAP processors online to the LPAR each hour.

X-axis

HOUR – Hour of Day

Query Defaults

Modifications

The following modifications can be made to the RMFZAC query: