A Task-Oriented Framework for Automatic Service Composition

Syntactic composition languages like BPEL WSCI are insufficient for end users because of many technical details. In order to reduce the complexity of service composition, various approaches based on semantic technologies and planning techniques have been developed for automatic service composition. However these approaches have focused on composition aspects and have not facilitated users expressing their business demands easily and flexibly. In this paper, we propose a semantic framework, that automatically transforms high level task definitions into executable processes. The task language OWL-T is developed for defining complex tasks that capture users' business requirements. Semantic technologies are utilized to map such task templates into executable processes expressed in the BPEL language. An implemented prototype was developed for illustrating the practical potential of the framework.