| Orchestrated web service applications are highly distributed applications that accomplish business goals by executing services offered by partners. This dependance on partner services allows the development of more flexible, modular applications. For a classical distributed system, correctness can be ensured by statically checking the composition of the components that make up the system against properties of interest. However, in the case of web service applications, there are various conditions that make this type of analysis insufficient. For example, partners can be dynamically discovered, which means that we cannot create a definitive model of the system to analyze. Web service applications can also display new behaviour at execution time, so statically checked properties of the system may not hold throughout the system's lifetime.;The first milestone for achieving this goal is the creation of an adequate property specification language. This language must be expressive enough to capture the distributed, interactive, and message-driven nature of web service applications, but must also be amenable to efficient runtime monitoring. We propose Web Sequence Diagrams (W-SD), a language that, we feel, meets these criteria. Specifications expressed in W-SD permit the analysis of orchestrations involving multiple partners, from the point of view of the orchestrating service.;The second contribution of this thesis is the creation of an industrial-strength online runtime monitoring and recovery framework that is non-intrusive, supports the dynamic discovery of web services, deals with synchronous and asynchronous communication, as well as partner services implemented in different languages. Developers using this framework can specify and efficiently monitor a variety of temporal behaviour. If recovery is enabled, properties are monitored proactively, so this framework allows developers to effortlessly enable error recovery in applications being monitored.;The last contribution of this thesis is the development of recovery plans from runtime errors. Given an application path which led to a failure and a monitor which detected it, we have developed various techniques and optimizations that make recovery plan generation feasible in practice. For some of the violations, such plans essentially involve "going back" -- compensating the occurred actions until an alternative behaviour of the application is possible. For other violations, such plans include both "going back" and "re-planning" -- guiding the application towards a desired behaviour.;Due to these limitations of static analysis, this thesis concentrates on the dynamic analysis of web service applications, specifically, by monitoring runtime events. The goal of runtime monitoring is to check whether an application violates a given specification of its behaviour during its execution. The behaviour of the system can be specified in a number of ways, e.g., as a set of temporal properties, assertions or even scenarios. During execution, application events are intercepted and used to determine if the system is violating its specification. Moreover, monitoring the system as it runs provides a chance to recover from an error once a problem has been detected. This is critical in the domain of web service applications, as bugs are potentially exposed to millions of users before they are found/fixed. We present techniques to address several major challenges facing the creation of an industrial-strength runtime monitoring and recovery framework for web service applications. |
