Applying Probabilistic Model Checking to Service Discovery Framework

Abstract

The complex requirements of E-commerce application call for selecting a set of Web services to reuse theirs business logics, where Service Oriented Architecture (SOA) provides a promising solution to the problem of cross-platform services integration. To this purpose, how to discovery services is a key to support the quick services composition, which has been a challenging task in Web application engineering. However, the traditional approaches have limitations in recall ratio and precision ratio because the keyword and semantic annotation query modes are hard to verify the target Web service, such as interaction behaviors and control flows of composite service. In this paper, it proposes a probabilistic model checking based service discovery method. First, it focuses on service reliability and time constraints to formalize the probabilistic behaviors of service process. Second, the quantitative verification properties of user requirements are specified in the form of temporal logic formulae. Third, the service discovery is to verify service process model against expected properties for identifying candidate services. Finally, in order to reduce model checking tasks, the correlation based service recommendation is introduced to explore more suitable services for user. Our framework improves the efficiency of service discovery without changing any existing service infrastructures.