SOA: Testing and Self-Checking

The dynamic nature of service-oriented architectures poses new challenges to system validation. Traditional testing is unable to cope with certain aspects of a service–oriented system validation, essentially because of the impossibility to test all (often unforeseen) system’s configurations. On the other hand, run–time monitoring, while able to deal with the intrinsic dynamism and adaptiveness of a service–oriented system, are unable to provide confidence that a system will behave correctly before it is actually deployed. In this paper we discuss the role of testing and monitoring to validate a service–oriented system and how they can be combined to increase the confidence and reduce the cost of validation.