Service Reconfiguration Architecture Based on Probabilistic Modeling Checking

Service software deployed in E-commerce and finance fields needs working under 7*24 houses mode. If any failure occurs, service reconfiguration should be immediately executed to find appropriate services from candidates in order to guarantee the availability of core business. Thus, service software cries for an effective approach to constantly adjust its form for responding to varying user requirements and instable runtime environments. To this end, this paper proposes a probabilistic model checking-based Web service reconfiguration architecture. First, it proposes a predictive Web service monitoring approach based on probabilistic model checking. Second, it gives a Web service dynamic service selection approach which takes compatibility checking into account. The single-source service selection works to execute service replacement, while the multi-source service selection carries out service simulation. Third, it discusses a Web service dynamic reconfiguration verification approach where the Probabilistic Counterexample-Guided Abstraction Refinement (Probabilistic CEGAR) is introduced to alleviate the state space explosion problem.