Identifying CICS constraints

Major constraints on a CICS® system show themselves in the form of external symptoms, stress conditions and paging being the chief forms. This section describes these symptoms in some detail so that you can recognize them when your system has a performance problem, and know the ways in which CICS itself attempts to resolve various conditions.

The fundamental thing that has to be understood is that practically every symptom of poor performance arises in a system that is congested. For example, if there is a slowdown in DASD, transactions doing data set activity pile up: there are waits on strings; there are more transactions in the system, there is therefore a greater virtual storage demand; there is a greater real storage demand; there is paging; and, because there are more transactions in the system, the task dispatcher uses more processor power scanning the task chains. You then get task constraints, your MXT or transaction class limit is exceeded and adds to the processor overhead because of retries, and so on.

The result is that the system shows heavy use of all its resources, and this is the typical system stress. It does not mean that there is a problem with all of them; it means that there is a constraint that has yet to be found. To find the constraint, you have to find what is really affecting task life.

If current performance has been determined to be unacceptable, you need to identify the performance constraints (that is, the causes of the symptoms) so that they can be tuned. This section discusses these constraints in the following sections:

[[ Contents Previous Page | Next Page Index ]]