Monitoring Web Services: A Case Study

Monitoring Web services proved to be a challenge after the Apollo Group chose to service-enable applications as a strategy for integrating front-end and back-end systems at its University of Phoenix subsidiary. The online university is one of the largest accredited providers of higher education, with about 375,000 students.

Monitoring Web Services: Successful Integration, but New Challenges

The Apollo Group wanted a mix of technologies at its University of Phoenix business to interoperate. Turning existing software—including Oracle, PeopleSoft, financial, in-house customer relationship management (CRM), and custom-built Java and .NET applications—into Web services achieved this goal.

But in the process—as the IT team responsible for integration created more and more Web services from existing applications—Web service management problems arose. While the team's traditional IT management tools for monitoring the experience of end users reported that one application was performing well, the Help Desk was receiving tickets from students about performance issues. The resulting audit found missing transactions. In addition, application and service owners had trouble determining who was calling their applications.

In other words, the new Web services were a black box, whose workings were not understood or accessible, so that the IT team couldn't monitor, track, and troubleshoot services performance issues. As a result, the team decided to find a tool to open the black box.

Monitoring Web Services—with Progress® Actional® for SOA Operations

The Apollo Group/University of Phoenix IT team chose Progress Actional for SOA Operations to address its issues and monitor over 70 Web services in production. With Actional, the team was successful.

Actional for SOA Operations allowed the IT team to automatically detect problems that were previously reported by students and staff. The team set up policies in the software for monitoring response times and faults and reporting on the results via email. In creating the policies, they set thresholds below the problem point. If the average response time or the percentage of faults goes up, they get alerts and are able to tackle potential problems proactively—before users see the impact.

In addition, applications owners have personalized portals, where they can view how their particular services are executing and even drill down to get views by operation and app server. They can also get reports on their services' performance by the hour, day, week, and quarter. These capabilities help IT staff pinpoint what and when something occurs with the services. As a result, they have reduced mean-time-to-resolution of problems by 70%.

Finally, because Actional for SOA Operations automatically discovers and maps what's going on in an integrated application consisting of diverse Web services end-to-end, it revealed that a service created for one application was, over time, being used by ten different clients. In other words, the product provided true visibility into what was actually going on—versus what is ideally going on, by design—in the organization's integrated Web services environment. This is critical for accurately enforcing security (for example, for ensuring that only authorized users can access key information) and for planning system capacity.

For More Information on Web Service Monitoring

Find out more about the Apollo Group/University of Phoenix Web service management solution and its benefits by registering to watch the on-demand webinar: Operating Your SOA—Are You Up to the Challenge?