As a major public, nonprofit health insurance fund, Socialistische Mutualiteiten (NVSM) knows well the mission-critical value of having a reliable, secure, and responsive IT system. This is one the reasons we had one of the first mainframe sites in Belgium. Today, 95% of our code and business logic reside on the mainframe. The requests come to the mainframe through the company’s mobile app 24x7 from doctors, patients and hospitals. Our service level objective is to reimburse patients within 24 hours – a commitment that goes a long way toward determining how we are viewed by our customers. In 2020, many regulations were changed each month to ensure customer needs were met during the COVID pandemic. With the evolution of regulations and healthcare guidelines for COVID, more customers found their way to our mobile application, driving the mainframe workload to over 170K CICS transactions a day. This shift required our relatively small mainframe staff to react quickly to the changes in a manner that demonstrated both business and IT agility.
The importance of being responsive and ensuring systems were up, scaling and responsive to customers has never been higher. We leveraged new techniques to ensure system resiliency and reliability. Traditionally, the mainframe monitoring is primarily with static thresholds. The static thresholds were often triggered and ignored as they were determined to be normal. On the other hand, sometimes there were no static thresholds to detect specific conditions. A specific example illustrates how static thresholds are a challenge. Let’s say the CPU threshold was set at 90%. On Monday mornings, CPU is always 90% or higher so the threshold is always reached, while on Friday afternoons the threshold is never reached.
We have a very large Broadcom stack and so engaged with Broadcom to discuss our needs. I learned that Broadcom included Mainframe Operational Intelligence with CA SYSVIEW Performance Management. That got my attention! Why? Because Mainframe Operational Intelligence (MOI) detects anomalies in system performance metrics, then dynamically raises alarms providing our team with insights which were previously hidden. More and more we have a special focus on what is abnormal. With visibility into what is abnormal, we then seek to understand what is driving the behavior and take the necessary action. We partnered with Broadcom to ensure the algorithms revealed meaningful and actionable insights.
Here are two specific business cases where we were able to capitalize on new insights enabled by Broadcom’s solutions.
With one of our applications it is normal to have 10 calls in a minute. MOI raised an alarm on the volume of transactions. Digging into SYSVIEW, we observed that 200 transactions were called from the same workstation, same user id. The user was leveraging 3270 emulation screen execution macros based on a file, and the emulation will execute many times. This is something we don’t want to support, so we helped the user find another path to meet their needs. Without MOI we would never have been notified as the view we had of the data was consolidated into a one-hour interval, and from the one-hour perspective, the abnormal activity was not noticeable.
Another case illustrates how hard it is when you must attempt to monitor activity that is not happening. We use MQ MFD to file transfer from Mainframe to distributed servers at a rate of anywhere from 0.5 to 1.5 transfers a minute. How do we monitor peaks when the rate drops to zero? Of course, there are rules that can be written such that over a period of time the transfer rate is observed to be zero. However, MOI makes it much easier. With MOI, when activity drops to zero, the system sends an alarm. Before MOI, we were limited to reading the log at the end of each hour-long period, so it might be a full hour before we noticed a problem. With MOI we get notified in less than 5 minutes. That makes a tremendous difference!
With the insights provided by MOI, we are in a much better position to continue delivering the high qualities of service that our business and our customers demand.
To learn more about Broadcom AIOps visit here.
Like what you read? Join our Mainframe Insights group to collaborate and ideate with us as we grow our Mainframe ecosystem together: https://www.linkedin.com/groups/9053158/