The CICSPlex Starburst Effect on Application Performance Measurement
When it comes to measuring CICS regions for application performance, analysts need to look at the CICSPlex as a whole as opposed to only looking at the performance of one or a few particular CICS regions.
Increased access by mobile devices to mainframe data via CICS transactions can create a starburst effect. This can be understood as a transaction executing in one CICS region causing a chain reaction of activity in many other CICS regions. In some cases, it’s possible one transaction could initiate upwards of 50 transactions.
As is standard, performance analysts continue to focus their measurement efforts on singular CICS regions without thinking about how activity in one region affects performance in other regions. Performance analysts aren’t looking at the CICSPlex to see where performance is poor, so what happens in other regions or how regions are interrelated remains uncharted territory.
How CICSPlex Affects Application Performance and End-User Satisfaction
Let me provide an example of why it’s important to find a way to measure the CICSPlex to solve an application performance issue, as opposed to only looking at one or a few CICS regions.
When you open your phone’s bank app to check your account balance or transfer money, the app accesses several to many distributed web servers and application servers. The final application server sends a request to a CICS region. When the CICS region picks up that request a transaction is initiated, and it initiates other transactions in other CICS regions.
If this business transaction has a service level agreement (SLA) on it, when the end user taps the button on their smartphone to view their account balance or transfer money, the bank expects the task will be accomplished and the user will be notified of the outcome within a certain timeframe. Let’s say it’s an SLA of three seconds, but the transaction takes two or three seconds longer than expected.
On the mainframe, performance analysts might look at an individual CICS region to attempt to determine the issue; however, if they fail to take into account how all of the transactions executing in the CICSPlex are making the application slower—a few hundredths of a second here and a tenth of second there—they quickly have an unhappy customer on their hands.
Focusing on the individual regions leads analysts to conclude that there may be no problem on the mainframe, as performance looks slightly above average. They won’t be able to piece together a thorough explanation for what’s hindering the transaction. As a result, the transaction will continue exceeding its SLA, application performance will decline and customers will continue to become unhappy.
As emerging technologies of engagement continue to increase the frequency and depth at which they access data in CICS regions, analysts must look at the CICSPlex performance to deduce where there might exist a starburst effect causing application performance issues. Attacking the application as a whole will ensure applications are running as efficiently as they’re designed to, thereby meeting and exceeding customer expectations.
Latest posts by Spencer Hallman (see all)
- Where Was the Mainframe at Big Blue’s InterConnect 2017? - March 27, 2017
- How to Hot-Rod Your Mainframe Applications and Save CPU - December 8, 2016
- The CICSPlex Starburst Effect on Application Performance Measurement - June 21, 2016