An Approach For Semantic Web Services Dynamic Composition Based On HP2P

Web Service (WS), as a set of de facto standard solution to establish Service Oriented Architecture (SOA), is very suitable for the integration of business applications. But the capability of single WS is very limited, so creating some new WSs for the complex business requirements by composing the existing WSs becomes the active research field in WS. Currently, Web Services Business Process Execution Language (WS-BPEL), providing the capability of describing business processes and business interaction protocols, is widely applied in composing WSs. And to improve the efficiency of service discovery in the composition, Semantic Web is introduced into WS, calling Semantic Web Service (SWS). SWS enables us to discover, invoke and composite WSs in an automatic way. With rapid growth of applications of WSs, the WSs composition systems based on the centralized management suffer from low availability, low accuracy and poor performance. However, the P2P Computing can solve the problems of WSs:single point of failure and network bottleneck.Based on WSs, Semantic Web, SWS, WS-BPEL and P2P technologies, this paper delivers a detailed approach for the dynamic Web Services composition in HP2P environment (HP2P4WS for short). HP2P4WS grounds on an OntologyDAG-based SWSs matchmaking and composition method. The WSs are composed by means of the Abstract Process Template (APT), which describes the interaction and business protocols among different WSs. To make semantic capabilities unambiguously interpreted and automatically used, we present a representation of ontologyDAG, based on labeling a spanning tree with numeric interval encoding, and storing redundancy concept information in an index table. To enhance the efficiency of service searching, an improved two-layer searching algorithm based on a restrictive finger table is proposed. Meanwhile, semantics is also brought into communication of message routing. Then the qualified WSs can be dynamically selected and bound into APT. Finally the HP2P4WS framework and working processes are introduced in detail to demonstrate the effectiveness of the approach.