Research On Semantic Service Discovery And Service Composition For Pervasive Computing

Pervasive computing is the integration of cyberspace and physical space. In this incorporated space, users can obtain services non-intrusively anytime and anywhere. The services in a pervasive computing environment are provided by a lot of autonomous, heterogeneous and distributed devices. So how to coordinate so many devices by software system is becoming a hot pervasive computing research topic. Service Oriented Architecture (SOA) is to be used to successfully as the framework for pervasive computing system construction. Because service discovery using interface or key-word does not meet the requirement of autonomous, heterogeneous and distributed pervasive computing environment, semantic-based service discovery becomes a hot research topic.In pervasive computing environment, a single service is very difficult to meet the user's task needs, so its need to use service composition to complete the complex tasks. As a highly dynamic environment, the services may join into or disappear from the environment. During the execution of a composition service, if one component service fails or becomes overloaded, a mechanism is needed to ensure that the running process is not interrupted and the failed service is quickly and efficiently replaced. This paper solves the problems faced by pervasive computing by semantic-based service discovery, conversation-based service composition and service composition adaptation. The main contents and contributions of our work are as following:(1) With the requirement of service discovery and service composition in pervasive computing environment, a Pervasive Computing Semantic Service Model (PSSM) is designed. The model describes a pervasive service by using service capability, service functional parameter, QoS parameters context parameter and service conversation.(2) By anglicizing the requirement of semantic service discovery of pervasive computing, a Three-layer Pervasive Semantic Service Matching Algorithm (TPSSMA) is given. The algorithm completes service discovery by matching service capability, input/output parameters and QoS parameters where the service capability and input/output parameters are described by semantic description.(3) By anglicizing the requirement of service composition of pervasive computing, pervasive service behavior is modeled to finite state machine and the Conversation-based Semantic Service Composition for Pervasive Computing Environment (CSSCPC) is given. Meanwhile by comparing the input/output parameters and QoS parameters, the most fitted service composition is chosen.(4) In pervasive computing environment, resources, equipments and users are highly dynamic, so service faults may be occurred. The Service Composition Adjustment Algorithm (SCAA) and Service Composition Reconstruction Algorithm (SCRA) are given.In a word, researches on semantic service matching, conversation-based service composition and service composition adaptation algorithm in pervasive computing environment which combination of semantic Web technology, service computing, finite state machine and pervasive computing, is a new method applying on pervasive service matching and pervasive service composition. The method improves service discovery precision and success rate of service composition because it fully consider the features of autonomous, heterogeneous and dynamic pervasive computing environment. It is really valuable to research and develop pervasive semantic service discovery and pervasive service composition system in order to develope pervasive computing research.This work is partially supported by the Ministry of Education Technology Research Key Foundation of China under grant (NO. 104086).