CPExpert DASD Component

Overview

In its MVS Performance Notebook, IBM stated that,

“over 75% of the problems reported to the IBM Washington System Center can be traced to some kind of I/O contention. Channel loading, control unit or device contention, data set placement, paging configurations, and shared DASD are the major culprits...”

Database Administrators or Analysts have time to address only the most serious problems.

The DASD Component of CPExpert is designed to identify the devices with the most serious DASD problems, and to provide the information necessary to solve the problems.

The DASD Component processes information from SMF (contained in your MXG, MICS, or SAS-ITRM performance database) to build a model of your I/O configuration. This configuration model is used by CPExpert to assess DASD performance problems.

The DASD Component includes an extensive user manual that...

  1. Indicates the likely effect of each finding on DASD performance
  2. Explains why the finding was made
  3. Suggests alternatives to correct the problem
  4. Provides specific references in IBM documents where the problem is described

Sample CPExpert DASD Output

Details

The DASD Component performs several types of analysis:

Basic Analysis… identifies the devices having the worst I/O problems, identifies the specific problems, and suggests ways to correct the DASD problems.

Shared DASD analysis… automatically detects when DASD problems might be caused by shared DASD. The DASD Component then analyzes data from all systems in your performance database to identify the systems that share the DASD devices.

Application analysis… can perform an analysis of the applications using your DASD configuration. The DASD Component uses a proprietary modification to MXG or MICS software to relate the DASD problems to specific workload categories (such as CICS or TSO).

VSAM analysis… investigates data sets that are responsible for more than 70% of I/O activity at some sides. Adjusting a few parameters or correcting common problems often can result in significantly improved performance (IBM benchmarks show as much as 90% performance improvement resulting from some simple changes).

Data set response analysis… can optionally identify periods in which specific data sets experienced DASD response problems. This feature is particularly important in a System Managed Storage (SMS) environment, since SMS manages where data sets are placed.

Components