An approach to optimize composite web services execution in a distributed servers cluster

With the growing availability of software applications as web services, web service composition is an emerging paradigm for enabling application integration within and across organizational boundaries. This paper presents a novel approach to partition a composite web service written as a BPEL program into a equivalent set of decentralized processes and deploy them into a predicable intranet based web servers cluster. It gives a revised genetic algorithm to partition a BPEL program represented as a program dependence graph and generate an optimized execution setup under a managed web servers cluster, with the goal of minimizing execution time of the process. Some empirical experiments and evaluations show and prove the efficiency of our approach.