A Data Flow Optimization Based Approach for BPEL Processes Partition

Abstract

Business agility usually requires dynamic business process management. For example, business process outsourcing (BPO), often needs partition an existing BPEL- described business process into a number of sub-processes, which can be run at different business partners' in a decentralized way. However, how to effectively partition BPEL processes is still a challenging issue. In this paper, we propose a data flow optimization based BPEL process partitioning approach. Firstly, the messages used in the BPEL processes are transformed into asynchronously transmitted variables. Then, the transformed BPEL process is represented as a parallel flow graph (PFG), and is optimized through data flow analysis. Finally, the business process is partitioned based on the optimized PFG. Experimental results have demonstrated that the workload of message transmission and the response time are all effectively reduced.