How do you pinpoint a performance problem in an enterprise-wide application that begins with a browser interface, passes through several servers, hops on to the mainframe for in-depth data and processing, and then makes its way back, stopping by another server or two, before showing up with the information the user requested, several minutes too late?

Like many things in life, it begins with admitting you have a problem. After all, many organizations that have mainframes have done so much of their new application development on distributed systems that they've forgotten where those applications go for their most important data and processing: the mainframe.

But when things do go wrong and the finger-pointing begins, it is too easy to become an illustration of that old saying, "when you point your finger there are three more pointing back at you" (referring to the three fingers in your hand that aren't being used to point, of course). At this point, problem solving gives way to a political blame game, while users and customers look to get results elsewhere.

Speaking as a mainframer, I'd much rather get results than play politics. So, as you can imagine, I'm happy that we've brought our distributed CA Wily APM (Application Performance Management) solution together with our mainframe CA SYSVIEW® product to provide a true end-to-end, enterprise-wide application performance management solution, as described in today's announcement of CA's Cross Enterprise APM.

Now, the mainframer and the distributed techie can sit down together at the same graphical dashboard, trace the activity of a transaction across its entire path to figure out where the problem is occurring, and then drill-down to the details on the platform and subsystem in question - for example, see the specifics of a particular CICS transaction's execution. They can even pull up application details such as the line of code or SQL that's causing the problem if they're using AQM's TRILOGexpert TriTune^®.

As an added bonus, this looks to me like an opportunity to help our less mainframe-aware colleagues learn about the role the mainframe plays in all their key applications that they've been assuming were only distributed. After all, any other view of production computing would be pointless!