IBM Information Management System (IMS) - Group home

Manage IMS performance spikes

By Jim Porell posted Fri June 12, 2020 09:23 AM

"Why is it taking so much time?" is heard from many IMS administrators lately. The answer is compounded by working from home, where they don't have access to what they were using before. Regardless, transaction volumes change for many reasons, including online sales, need to access banking transactions from home, package tracking, etc. Making matters worse,  customers complain about elongated transaction times. Longer than normal transaction times are frustrating.  For retail, this could lead to lost sales as a customer has choices from whom they buy. In banking and package tracking, customers may repeatedly ask the same questions, resulting in multiple duplicate transactions queuing up and making the situation worse.

Customizable utilization reports

IMS Performance Analyzer for z/OS is built to handle these situations. IMS PA measures the time spent in IMS as gathered from a variety of sources. The picture provides a view of some of the components and measurement points available to IMS PA.
IMS Transaction time 
Through the data that's available, IMS PA has a number of fixed and fully customizable reports that can show transaction data.  Highlighting the transactions that are "out of the norm" reduces the time and effort on the specific issues at hand.

Deadlock detection

Typically, the ailing transaction might be encountering a System Deadlock. Using IMS PA and the Deadlock report, administrators get a comprehensive analysis of deadlock events. These reports help with tuning applications and adjusting their scheduling to avoid this expensive overhead.  That overhead contributes to transactions waiting for elongated periods.

Buffer analysis

Memory, storage and processor time are critical to any transaction. In addition to the CPU time analysis mentioned early, reports on resources and buffer pools are also available. Inefficient usage of memory easily impacts transaction times. Remediation will be identified enabling more transactions to run smoother.

Building historical transaction index records

As a system, z/OS captures myriads of data. IMS PA leverages that data to build transaction index records that capture performance metrics in a single log record that is easily customizable to include external subsystem and database activity. These log records should be compared to historical records to ensure things are moving well or quickly identify hot spots.

Keep your IMS system running smoothly

Regardless of the type of IMS transactions that are running, external and internal events change the patterns of historical success. Using IMS PA will keep your systems running smoothly and your consumers satisfied with their experiences. To learn more about IMS PA, visit: