Abstract-Webservice-business process execution Language (WS-BPEL) is a promising language for describing a web service composition. Unfortunately, WS-BPEL lacks expressive power of formal semantics to support verifiability. Consequently, timed trace theory has been used to assure the correctness and reliability of the web service composition with timing constraints. Moreover, since most of business processes are more complicated, the cost of verification is high. It is likely that the state explosion problem will take place. In order to tackle this problem, in this paper a partial order reduction algorithm for timed trace theoretic verification has been applied to minimize the number of states. Experimenting with a tourism information system, the proposed approach expresses its effectiveness.Index Terms-Partial order reduction, timed trace theory, web service composition.
I. INTRODUCTIONWeb service composition is an ability to integrate existing web services together to fulfill the practical requirements. Presently, WS-BPEL (Web Service-Business Process Execution Language) [1] has been introduced to composite web services. Since creating composite web services is more complex task, formal verification of WS-BPEL is essential to guarantee the correctness and reliability of web service compositions with timing constraints [2]. However, WS-BPEL has some defect on verification.Recently, a number of research efforts on the verification of the web service composition have been activated including [3]- [8]. Even though all of them are successful in unraveling weaknesses of verification of web service composition, they do not consider timing constraints of business processes. Therefore, [9]- [13] have been proposed formal semantics for verifying WS-BPEL with timing constraints. Whereas they are not able to detect safety and timing failures, timed trace theoretic verification proposed in [14] does. Using this approach, a specification and implementations are modeled by time Petri net [15] and conformance relation between them is checked.However, since practically most of businesses processes not only involve functionality but also involve the timing constraints, they are quite complicated interactions. The cost of verification is rather high. It is likely that the number of states may explode or the state explosion problem will occur. Hence, a partial order reduction algorithm for timed trace theoretic verification [16] has been applied to reduce the number of states while verifying the web service composition with timing constraints. This paper is organized as follows. Timed trace theoretic verification and its partial order reduction algorithm are described in the Section II. How to transform WS-BPEL into time Petri net is explained in the Section III. Then, experimenting with a case study is demonstrated in the Section IV. Finally, the conclusion is given in the Section V.
II. PARTIAL ORDER REDUCTION FOR TIMED TRACE THEORETIC VERIFICATIONSince our approach relies on the framework of partial order reduction for timed t...