CICS Transaction Timing Analysis by High Level Components

Examples of CICS transaction response time analysis starting from high-level categories such as CPU, I/O Wait and Other Wait.

More CICS Transactions Videos

1. CICS Timing Metrics – Overview
2. Transaction Response Time Analysis
3. Timing Analysis by High-Level Components
4. Response Time Analysis Scenario 1 Values Differing by System ID
5. Integrating WLM Performance Index and CICS Transaction Analysis
6. CICS Web Services Metrics
7. Overview of Non-timing Transaction Metrics

Video Transcript

Okay. So another approach we might take as we’re analyzing timing metrics could be to look at the profiles of the top transactions, kind of grouped by those major buckets, major high-level categories. So for example, for I/O weight, you can see most of the transactions in the Top 20 volume here have Wait for Inbound Socket I/O, or Wait for Control at the end of MRO. If we look at other weight, that’s the other kind of primary category of suspend time. We actually have one transaction here that’s kind of compressing the scale for everybody else. So let’s just remove that one.

And when we do that, you can see here that TCB Switch Wait Time is a primary contributor to many of these transactions, the other wait time. And then there are other measurable components that other transactions have here as well. Now, CPU time per Tran typically gets a lot of attention in CICS performance analysis. So we talked earlier about the many types of TCBs, the CICS Dispatcher manages. And so here we see the CPU consumption on each of those types of TCBs captured at the transaction level.

So, for example, here in high transaction volume environments, you want to keep an eye out for significant changes in CPU per tran over time, for example, before and after the implementation of an application or release. So typically, that kind of comparison would cover the weekend that the release went in. But since I just have a week in my demo database here, I’ll just compare to the previous day. And that shows a minimal change.

We commented earlier that some sites have to carefully manage the utilization of the QR TCB. And if that’s the case, then you probably would want to look at that metric by transaction. And let’s also sort it by that as well. So here we see the transactions that are consuming the most CPU time. And actually let me go ahead and change the title here, and then we’ll go ahead and put that in our dashboard. And if the QR TCB is more constrained at certain times than others then maybe we would want to view that consumption over time.