Web Services Performance

How do we measure Web services performance in runtime against what is supposed to be happening (against, that is, what was designed)? This is one of the key issues involved in optimizing the SOA environment. Many of us are familiar with scenarios in which the architectural diagram or network diagram is obsolete, it seems, nearly as fast as it is published.

Runtime service control, then, is about answering questions like: what is actually happening in production? Who's actually using our services? What we often find is that production does in fact look different than what was originally intended.

Measuring and Controlling Web Services Performance

So it's always good to get fact-based confirmation of what's happening in production. Consider the various dependencies among applications and services. The registry can help at design-time to answer the question, "Who is using this service?" The service owners, consumers and dependencies will be listed in the registry. But it will also be important to confirm, for example, what happens if a service is versioned? What other services, processes, SLAs will be impacted? Can downtime be tolerated?

How to Be Certain About Web Services Performance?

With respect to policy enforcement, IT needs to pay attention, specifically to service-level enforcement. In order to do this, IT has to have the means to be certain that SLAs are indeed being enforced. How to confirm, for example, that all is being forced through an IDM or message broker? These technical capabilities will be addressed in the next tutorial on governance solution technologies.

For More Information

Discover the secrets of synchronizing Web services performance in runtime with services in design-time. Download the free webinar, Optimizing the SOA Lifecycle from Design Time to Runtime with Comprehensive Governance