In this age of hybrid IT, CICS applications serve as the critical back-end to many web, cloud, or mobile applications. Those who manage these hybrid applications are able to accurately monitor what’s going on outside the mainframe with relative ease. Likewise, the IT group can get precise data about what is happening inside the mainframe. The difficulty is understanding how what is happening off the mainframe is driving transactions on the mainframe. Without a complete view of hybrid app-mainframe interaction, mainframe optimization is almost impossible to perform with any precision.
What is apparent is that mainframe transaction volumes experience spikes, but certainty about where it’s coming from and why is elusive. A poll of attendees in a recent IBM Systems Media webinar shows how unaware organizations are of how hybrid applications interact with the mainframe.
One of the latest sources of mainframe transaction volume is Robotic Process Automation (RPA). As RPA adoption goes mainstream in the enterprise, some IT groups are trying to understand the impact it has on the mainframe. HostBridge is actively working with several enterprises that suspect RPA activity is the root cause of some operational issues. This scrutiny doesn’t occur in the name of trying to shut down “shadow” IT automation efforts. Instead, it’s part of an effort to optimize mainframe automations. Integration analytics provide insights into all these interactions, helping IT organizations holistically manage hybrid IT applications and integrations, from end-to-end.
Integration analytics are what provides visibility into the myriad and often-complex interactions involving the mainframe. By blending transaction data, analysis processes, and dashboards, integration analytics fill the performance data gap between off-host and on-host applications. The process is able to correlate non-mainframe service requests to actual work done inside CICS to process those requests. It allows the enterprise to gain a precise and real-time view of how activity off the mainframe drives activity on the mainframe. This visibility is where the power of integration analytics comes in: it enables better management of hybrid IT applications and infrastructure.
The foundation of integration analytics are SMF 110 records. The analysis of this CICS transaction data can provide the enterprise with a clear picture of cause and effect. When necessary, the HostBridge Transaction Analytics Connector (HTAC) software can extract origin data and metadata from external requests. HTAC then annotates the SMF records with this data. The detail it provides enable diagnosing, remediating, or optimizing performance issues.
Why are integration analytics necessary?
Sometimes, non-mainframe applications are not “good citizens” when it comes to mainframe interaction. There’s no malice here, just ignorance. Those who make decisions about mainframe access methods and hybrid IT architecture often don’t understand the impact of their integration choices on the mainframe. They scrape data from the UI – because it’s what they know to do – instead of using an API. They don’t understand the impact of issuing unnecessary preflight requests to the mainframe from middle-tier application servers. Or, they send one request per CICS transaction instead of orchestrating an entire transaction sequence. Over time, these inefficiencies build up like plaque in the arteries that connect the mainframe to the hybrid IT infrastructure.
Like real life, too much plaque buildup in the integration arteries results in performance issues. Sometimes, the result is the premature death of otherwise valuable CICS applications. (It’s almost always the mainframe application that takes the rap for poor performance or integration difficulties). Using integration analytics is the functional equivalent of a vascular ultrasound scan. It tells you if there’s a problem, where it is, and what you can do about it. These use cases show the value of integration analytics:
- An auto insurer was experiencing unexplained spikes on the mainframe when clients or agents added vehicles to existing policies. This insurer uses a third-party CICS application for policy management, and it had no visibility into how the application consumed mainframe cycles. With help from HostBridge this insurer began using integration analytics. The Splunk dashboards HostBridge developed exposed what was happening inside the policy application: millions of unnecessary instructions were running. With this analytic data, the insurer was able to get the vendor to make applications changes to improve the efficiency of this key business process.
- A government agency was trying to understand the source of unexplained CICS workload spikes. These spikes were driving up processing costs and creating a performance bottleneck. Using integration analytics, the HostBridge team was able to trace the spikes to their point of origin off the mainframe. The analytics revealed that application servers were using a single-request to run a single transaction on the mainframe. The solution was to use a far more efficient approach of orchestrating an entire sequence of mainframe transactions from a single application server call.
- A global technology supply chain vendor was experiencing asymmetric increases in mainframe transaction volumes. The analytics confirmed the suspicions of the IT group. Years of screen scraping Excel macros and now RPA bots were driving the increases. As these automations proliferate through the enterprise, the impact to the mainframe becomes more severe. Now, a series of custom Splunk dashboards display the integration analytic data. With them, the IT group can identify the “heavy hitters.” In many cases, the resolutions are simple, once the source of the increase is known.
Who needs Integration Analytics?
Any enterprise whose CICS applications participate as part of hybrid IT needs to leverage integration analytics. This means almost every enterprise, as mainframes are by default participants in hybrid IT strategies. Signs that indicate integration analytics can help include:
- Unexplained or uncharacteristic CICS transaction spikes
- Unexpected performance bottlenecks
- No or limited visibility into what is driving CICS workload
- Suspected inefficiencies in mainframe applications or integration practices
HostBridge can help you leverage integration analytics.