When you think of the typical mainframe user, who comes to mind? You may not realize that this person is actually a mom doing home banking, a salesperson booking a flight, or a college student ordering school supplies.
The mainframe is a critical component of today’s transaction-oriented, multi-tier business applications. Requirements are now driven by a large population of users who care only about how long it takes to complete their transactions. They aren’t concerned with the underlying infrastructure handling those transactions.
These users are now in the driver’s seat. To meet their needs, IT must monitor not only the performance of mainframe components but also the performance of transactions from the user’s point of view. Additionally, IT must gain visibility across the entire path taken by transactions as they traverse the various application components, both mainframe and distributed. A major challenge is understanding how each individual component that supports the user can be kept current.
IT organizations that are tuned into this new reality are evolving their application performance management strategies accordingly. They know that monitoring the user experience is essential to gauging the overall health of applications. And they’re working to attain end-to-end visibility into performance, all the way from the user at the front end, who is relying on information from the mainframe at the back end.
The paths transactions take today are anything but direct. When a customer submits an order on a company’s Website, the transaction may involve inventory and order-entry systems on the mainframe, as well as distributed components such as Web servers, application servers, and middleware. Many applications also invoke services, such as credit card verification, provided by outside suppliers. Factor in the growing number of people using Web applications and you have a significant challenge in monitoring and maintaining the performance levels the business requires.
Mainframe performance management tools have evolved over the years into sophisticated solutions that enable you to ensure the mainframe and all its components—CICS, IMS, DB2, and WebSphere MQ—are meeting performance requirements. With these tools, the staff can closely monitor what’s happening within the mainframe.
Today, however, it’s important to expand the scope of application performance management. In addition to knowing what’s happening at the back end, you need an even greater understanding of what each user is experiencing. That visibility must include such elements as transaction response times and the type of applications and information being accessed. You must factor that information in with any other metrics you use to understand true application performance.
Tools have evolved to monitor and manage performance across distributed infrastructure components. There are also tools for monitoring user experience, and still others for transaction tracing. These tools should be integrated. Otherwise, the staff grapples with multiple tools from multiple vendors and manually combines data from multiple sources to get broader visibility.
A solution that integrates the tools will provide a unified, holistic view from the user through service providers, into the infrastructure with its application layers and middleware, and all the way to the mainframe applications and databases. The solution should present user-experience metrics to enable IT to spot issues early and take action before performance degrades. It must also provide a comprehensive view of application dependencies so the staff can prioritize actions. So, if one problem is preventing employees from accessing information about company health benefits but another problem has slowed order entry on the Website to unacceptable levels, it’s clear the problem affecting the Website and end users is the first to fix.
The concept of application performance management is pretty simple: You detect and resolve performance issues before they affect users. But the number of people accessing applications on the mainframe has skyrocketed, and performance management is now a highly complex discipline. Going forward, any successful strategy must account for the fact that users are in the driver’s seat when it comes to judging performance. IT needs the visibility across the enterprise to ensure that users have an underlying “engine” with the speed and performance they’re seeking.