In recent years, processor speed and VSE/ESA's ability to run more jobs concurrently have vastly increased. But these innovations only go so far; since VSE/ESA still uses primitive partition balancing techniques to manage CPU resources. As you know too well, CPU is a major bottleneck in VSE systems. When one job or application dominates available CPU resources, you get sluggish response times and long run-times.
Finally, there is a solution to managing the CPU bottleneck: Barnard Software's OPTI-WORKLOAD. This revolutionary product introduces the concept of performance groups to VSE/ESA. OPTI-WORKLOAD's intelligent CPU management results in more equitable allocation of resources and greater overall system throughput.
The priority of static and dynamic class partitions can be set and changed by using the AR PRTY command. When a dynamic class contains more than one allocated dynamic partition, the partitions within the dynamic class are balanced (time sliced). The time slice value can be modified via the MSECS command.
The VSE/ESA partition balancing routine inspects the CPU time used by each partition in the balancing group and decides how to rearrange the priority of the partitions. VSE/ESA 2.1 running the Turbo dispatcher offers a new balancing algorithm. If partition balancing is specified for static and dynamic classes (via equal signs in the PRTY command), static and dynamic partitions will receive the same time slice. Without the Turbo dispatcher, a dynamic class (with all its partitions) receives the same time slice as a static partition.
OPTI-WORKLOAD introduces the concept of performance groups to VSE/ESA 1.3 (and higher) systems. Performance groups are groups of static and/or dynamic class partitions with similar performance goals. Once a performance group has been identified, OPTI-WORKLOAD monitors the velocity-of-work being done by each partition and class. At specified intervals, the partition or class achieving the lowest velocity-of-work is given the highest dispatching priority of those partitions and classes in the performance group. This results in a more equitable allocation of resources and greater overall system throughput.
OPTI-WORKLOAD has special OPTIMIZE and VENDOR features designed specially for VSE/ESA 2.1. These features, when used in a VSE/ESA 2.1 environment, can reduce overall CPU usage by 5-15% or more. The OPTIMIZE feature improves the interface between the application program and the VSE/ESA 2.1 supervisor which results in reduced CPU usage. The VENDOR feature enables or disables various VSE/ESA 2.1 vendor interface calls within the supervisor. Disabling of the vendor exit calls, when they are not needed, can provide a dramatic reduction in CPU usage. Best of all, when using the new VSE/ESA 2.1 Turbo dispatcher, the CPU usage savings is all non-parallel CPU time. This allows the Turbo dispatcher to handle multiple CPUs more effectively.
OPTI-WORKLOAD provides a special CICS Workload Management feature. The CICS Workload Manager uses Velocity calculations to manage CICS transaction priorities. Proper CICS Workload Management results in dramatic improvements in CICS response times and throughput. The CICS Workload Manager uses Velocity calculations to increase the priority of CICS transactions with low velocity-of-work (or transactions 'starving' for resources) and to decrease the priority of CICS transactions with high velocity-of-work (or transactions monopolizing resources).
Velocity-of-work is calculated by monitoring the number of times a partition actively is using the CPU, actively waiting on I/O, delayed waiting on the CPU, and delayed waiting on other resources. Velocity is expressed as the percentage of the time needed resources are available when the partition requires them. Velocity is not a measure of the quality of the work being done, it is simply a measure of how much of the time resources needed by a partition are available to it.
Performance groups are used to identify partitions with similar workload characteristics and performance goals. The simplest performance group definition for most installations is the ONLINE group. For example, the ONLINE performance group for an installation running two production CICS partitions (class C), VTAM (F3), DB2 for VSE (F2) and a test CICS (F4) might appear as follows:
GROUP ONLINE C F3 F4 F5
ALWAYS-LO F4
OPTI-WORKLOAD normally manages performance groups by monitoring the velocity-of-work being done by each partition or class within the group. Performance groups can also be managed through traditional CPU and IO balancing techniques.
The CPU command can be used to inform OPTI-WORKLOAD that traditional CPU balancing should be used to manage the specified performance group. When using CPU balancing, the partition using the least CPU is given the highest dispatching priority in the next time slice.
The MINCPU command can be used to inform OPTI-WORKLOAD that the specified partition or class is to receive a minimum percentage of CPU. If the partition or class does not receive the specified percentage, it is given the highest priority within its performance group.
The MAXCPU command can be used to inform OPTI-WORKLOAD that the specified partition or class is to receive a maximum percentage of CPU. If the partition or class receives more CPU than the specified percentage, it is given the lowest priority in the system during the next time-slice.
The IO command can be used to inform OPTI-WORKLOAD that traditional I/O balancing should be used to manage the specified performance group. When using I/O balancing, the partition with the highest I/O rate is given the highest priority.
OPTI-WORKLOAD runs under VSE/ESA Version 1.3 or higher. The product uses 4K of 31-bit system GETVIS.
To manage your CPU bottleneck, call 1-800-251-OPTI today.
We are so confident that OPTI-WORKLOAD will improve your system's
performance that we offer a free, 30-day product evaluation.
© 1996 Barnard Software, Inc.
All names, products, and services mentioned are the
trademarks or registered trademarks of their respective
organizations.