A context-aware computing mediated dynamic service composition and reconfiguration for ubiquitous environment

Context-aware and customization services are expected in the forthcoming ubiquitous computing environment. In order to make SOA provide context-aware services, we propose a SOA-extended architecture, which integrates a Context-aware Decision Making system that can evaluate services for mediating dynamic service composition and reconfiguration according to the context information. In the Context-aware Decision Making system, a semantic multi-dimension model is defined for the context representation and inference. By using this context model, we explore an effective mechanism that can evaluate all services from two aspects: QoS and user preferences. To evaluate user preferences for services, a method based on Association Rules in data mining is defined. This evaluation result is a configuration that can guide SOA to select services that are most appropriate for the current context environment. Finally, a language called Abstract BPEL and a service broker middleware are included to assist our architecture to enable dynamic service composition and reconfiguration. In this paper, a case study with intelligent house is also presented for verifying the featured functionalities of our architecture.