Sysplex Calculator: How to Interpret Output
The output from the Sysplex Calculator job is referred to as the Sysplex Calculator Report. This report is broken in to four major sections in the following order: overview, disclaimer, summary and interval-by-interval details. Each section is explained below:
Overview and Disclaimer Sections
- release date of the tool
- customer name (read from parameter)
- Serial number of each machine being analyzed (only last four digits of serial number can be determined from the SMF70 record)
- Software MSU rating for each machine being analyzed
- List of LPARs for each machine being analyzed
- Footnote beside an LPAR name if an LPAR is being used as the MASTER timezone
- Footnote beside an LPAR name if an LPAR with general purpose CPs exists on a machine but no SMF data was provided
- legal disclaimers regarding the tool (always review your Sysplex Calculator reports with IBM to ensure validity)
Summary Section
Based upon the data analyzed, the Sysplex Calculator will make a determination about which Parallel Sysplex each machine belongs to, based on the 50 percent criterion contained within IBM's Parallel Sysplex aggregation rules. A machine is said to belong to a particular Parallel Sysplex only if 50% or more of the eligible operating system workload participates in the Parallel Sysplex. For each machine, the following are the three conclusions that the Sysplex Calculator may reach, based on analysis of the prime-shift weekday intervals:
- If (looking at an average over the 40 hour prime shift data in the input stream) 50% or more workload is generated by a particular Parallel Sysplex, then the Sysplex Calculator will conclude that the machine officially belongs to that Parallel Sysplex.
- If the average utilization show that a machine belongs to one Parallel Sysplex for part of the period and another Parallel sysplex for part of the period, then the Sysplex Calculator will conclude that the machine belongs to *mixed* Parallel Sysplex. This means that the machine cannot aggregate with other machines.
- If the average utilization shows that a machine belongs to no Parallel Sysplex, then the Sysplex Calculator will conclude that the machine belongs to *none* Parallel Sysplex. This means that the machine cannot aggregate with other machines.
Interval-by-Interval Section
There are seven header rows shown in the Interval-by-Interval section. The first header row shows the machine for each LPAR analyzed. The second header row shows the LPAR name for each LPAR analyzed. The third header row shows the SYSID name for the instance of z/OS running in each LPAR analyzed. The fourth header row shows the System Name for the instance of z/OS running in each LPAR analyzed (this is used to correlate with the System Name in the Coupling Facility Activity Reports). The fifth header row shows the the Parallel Sysplex name for each LPAR analyzed. The sixth header row shows the average utilization contribution (in %) of each LPAR's workload, as a percentage of workload on the machine over the 40 hour prime shift period. The seventh header row shows the average utilization contribution (in MSUs) of each LPAR's workload, over the 40 hour prime shift period.
For each interval, you will see an MSU value listed for each MVS-based LPAR analyzed. The MSU value represents the average MSUs consumed by that LPAR, during that interval.
Contact IBM
- Email IBM
- Find a Business Partner
-
Call IBM: 1-866-883-8901
Priority code: 101AS13W
Reference info
Parallel Sysplex Aggregation Frequently Asked Questions (2 Nov 2004) (112KB)
United States Parallel Sysplex Announcement Letter (2 Nov 2004) (15KB)
EMEA Parallel Sysplex Announcement Letter (2 Nov 2004) (13KB)