The situation

A financial institution, outsourcing its IT operations, has been a satisfied SMT customer for more than 15 years and so knew who to call when their night batch suddenly seemed to suffer from a lack of CPU resources.

There had been no changes to applications or business volumes. Nevertheless, the system seemed to be overloaded and the night batch continued to execute well into what should have been online opening hours.

The response

SMT Data’s consultants responded quickly and provided the customer with information showing the cause of the problem. This Information was passed on to the outsourcing provider immediately (the very same day).

The result

It turned out that a previous upgrade of the operating system had introduced an error, which periodically multiplied the system overhead (for reasons unknown). The vendor chose to circumvent the problem and offered a substantial free memory upgrade (which actually made the overhead peaks disappear again). The service provider offered the SMT customer several months of reduced billing (for the entire period, within which the error could possibly have occurred).

But . . .

Still, two questions remained. The underlying problem had not really been solved (no software had been corrected) so could the system overhead pop up again? And if so, could it be less serious and go unnoticed, slowly increasing the usage based billing? The answer is no, not when SMT Data’s periodic optimisation audits are deployed.