QoS Driven Semantic Based Grid Service Composition Using AND/OR Trees

Composition oriented service discovery is an important requirement and research challenge in Service Oriented Architectures (SOA), such as in the third generation of Grid. It provides added value services, more rapid application development and improved reusability of existing services. This paper proposes and formulates a composition approach where semantic information is used to determine Grid services dependencies in form of AND/OR tree associated with semantic QoS information, thus transforming the issue of service composition discovery to constrained AND/OR tree search problem. Different constraint forms and QoS aggregation patterns in such trees are analyzed, and some major constrained searching techniques that can be applied to such trees are discussed. Our findings show that AND/OR trees are expressive in addressing QoS - aware semantic Grid service composition and able to employ different discovery searching techniques for that purpose.