Modeling and performance evaluation of service choreography based on stochastic Petri net

Abstract

Web service composition refers to the creation of new services by combining functionalities provided by existing ones. Current web service choreography proposals, such as WSCI, WS-CDL and BPEL, provide notations for describing the message flows in Web service collaborations. The kernel of WSCI consists of simple communication primitives that may be combined using control-flow constructs expressing sequence, branching, parallelism, synchronization, etc. Many efforts have been made on functional formalization and property verification of WSCI-based service compositions. However, performance facet of WSCI-based service composition is yet to be given the importance it deserves. In this paper, we introduce a stochastic-Petri-net-based approach to predict the performance of web service choreographed using WSCI. This approach translates WSCI-based composition into general stochastic Petri net and obtains analytical estimation of process-completion-time based on the continuous-time Markov chain derived from the Petri net. In the case study, we use Monte-carlo experiments and confidence interval analysis to validate theoretical evaluations.