Integrated Service Adaptation

With more automation in inter-organizational supply chains and proliferation of Web services technology, the need for organizations to link their business services and processes is becoming increasingly important. In some cases, we need adapter to reconcile the communication incompatibilities between two processes. My thesis proposes a new framework for integrated service adaptation. For the control flow adaptation, we propose a structural analysis of patterns which we show is more efficient than existing approaches. Furthermore, we propose an algorithm based message pair analysis and integer programming to create an optimal adapter. For message adaptation, we identify a set of extendible message adaptation patterns to solve typical message mismatches. In addition we give an algorithm for generating new message adapter on the fly so as to integrate control flow considerations into message adaptation. Finally we show another algorithm to integrate individual message adaptation patterns with control flow adapters to create a complete adapter for two processes. We implement all these algorithms in a Java-based prototype system and show the advantages of our methods by performance experiments.