Enhancing the Testability of Web Services

Abstract

For the foreseeable future, testing will remain the mainstay of software quality assurance and measurement in all areas of software development, including Web services and service-oriented systems. In general, however, testing Web services is much more challenging than testing normal software applications, not because they are inherently more complex, but because of the limited control and access that users of Web services have over their development and deployment. Whereas the developers of normal applications, by definition, have full control over their application until release time, and thus, can subject them to all kinds of tests in various combinations (e.g., integration testing, system testing, regression testing, acceptance testing, etc.), users of Web services can often only test them at run-time after they have already been deployed and put into service. Moreover, users of Web services often have to share access to them with other concurrent users. In order to effectively test Web services under these conditions special measures and approaches need to be taken to enhance their testability. Right from the early phases of development, the testability of services needs to be taken into account and “designed into” services. In this chapter we consider these issues and with the aid of a case study we present a methodology that can be used to enhance the testability of Web services.