Enterprises that run core business applications on mainframes have, for decades, integrated those applications. Today’s normal is for the mainframe to function as a key part of a hybrid IT environment. The first challenge enterprises face is how to complete these integrations. When integrations are complete, a new challenge emerges: understanding how these integrated, hybrid applications are performing. It’s often relatively simple to get data that shows how individual components of composite applications are performing. More problematic, historically, is gaining a holistic set of analytics that reveal true end-to-end performance. What enterprises need are comprehensive integration analytics.
When the mainframe is part of a hybrid IT application, it’s important to monitor data flows, because mainframe applications are typically high-volume and mission-critical. Seemingly “small” changes integrations introduced at any layer of a hybrid IT application can have a major, ripple-through effect. A middle-tier integration approach introduces a few milliseconds of additional latency into a transaction. What is the end-user impact of that latency when multiplied across millions of transactions a day? An architect opts for an integration approach that runs one CICS transaction per call, instead of orchestrating a full sequence of transactions per call. What is the impact on general processor cycle consumption for high-volume transaction applications?
From our work with mainframe-using enterprises, we have seen three specific areas of interest emerge relating to integration analytics: end-to-end transaction tracking, application performance analytics, and RPA analysis.
End-to-End Transaction Tracking
An example illustrates the need for this type of integration analytics. When a bank customer uses a smartphone app to check account balances, what happens on the back-end, mainframe side of the transaction? Enterprise IT groups want to see the complete picture, understanding how smartphone app usage is driving mainframe transactions. They also want to know where performance bottlenecks may exist.
An example of end-to-end tracking comes from a public-sector HostBridge customer that was looking for a way to track unexplained CICS workload spikes to their point-of-origin off of the mainframe. These spikes were driving up processing costs, creating a performance bottleneck, and consuming cycles that prevented other workload from processing. Implementing an integration analytics process enabled this client to see complete, end-to-end, hybrid IT application metrics on their Splunk dashboards. Through integration analytics, this client discovered that orchestrating CICS transactions on the mainframe and making them available as callable web services would significantly reduce processor consumption.
Application Performance Analytics
Many IT organizations experience frustration at how little they are able to know about how the performance and operations of their mainframe applications. Performance details of precisely what things an application is doing to consume cycles is often hidden.
An auto insurer was seeing mainframe cycles spike each time a client would add a new auto to a policy. It uses a third-party policy management application and had no insight into how that application was consuming cycles. HostBridge developed a series of integration analytics Splunk dashboards for this insurer, providing previously hidden application performance details. These dashboards led to the discovery that adding a vehicle to a policy was causing 28 million, unnecessary host instructions to execute. The insurer was able to work with their application vendor to remediate the processing spike and reduce overhead.
Robotic Process Automation (RPA) analysis
A fairly recent area of interest involves understanding how RPA activity drives mainframe transaction volumes. Many enterprises are starting to see RPA activity that points to the mainframe scale up. As it does, some of these enterprises are experiencing performance aberrations and asymmetric increases in mainframe transaction volume. Integration analytics provides insights into which RPA bots are interacting with the mainframe, how they’re doing it, and what transactions they are triggering. Without integration analytics, RPA’s impact on the mainframe remains a mystery. With integration analytics, enterprises can optimize these interactions to improve performance and lower costs.
One enterprise was experiencing asymmetric increases in mainframe transaction volumes, as RPA bot activity proliferated through various business divisions. The IT team thought RPA activity was the source of the increase, suspicions that integration analytics confirmed. In one case, a single bot was triggering 7,000 screen-scraping transactions several times a day. Through the integration analytics dashboard, the IT group was able to see what this bot was doing. Insights from the integration analytics dashboard let the IT group displace these 7,000 screen-scraping transactions with a handful of far more efficient DB2 calls.
How to Enable Integration Analytics
Detailed mainframe application performance data is often the missing piece of the integration analytics puzzle. Adding this data to complete the picture is the HostBridge Transaction Analytics Connector (HTAC).
HTAC is software that runs under CICS and extracts data and metadata from each external request to run a transaction/program. HTAC delivers this information to Splunk, along with the standard information that CICS keeps track of for all transactions. Using the data and metadata HTAC extracts enables creating Splunk dashboards that correlate mainframe activity to non-mainframe activity. HostBridge provides services and consulting to build custom Splunk integration analytics dashboards. Through these dashboards, HTAC weaves together an end-to-end view of hybrid IT application performance. Customers get a complete view and the insights from it allow them to remediate and optimize the performance of these applications.
Prototype Your Own Integration Analytics Dashboard
HostBridge provides trial versions of HTAC, with support and services to help you build a custom, integration analytics dashboard. Contact us using the form below to learn how you can gain insights like the ones this post discusses.