An infrastructure for executing WS-BPEL workflows in a Cluster of Clusters

摘要

The use of heterogeneous distributed systems, such as Clusters of Clusters (CoC) and computational grids, enables the development of complex computational systems. However some of these systems require services to be easily composed and executed while maintaining the dependencies among them; both the services and the dependencies can be represented as workflows. Moreover, Web Services have been adopted in distributed systems, and WS-BPEL is the standard for Web Services composition. Accordingly, the objective of this work is to enable the execution of WS-BPEL workflows in a CoC. In order to achieve this execution, an extension to the WS-BPEL language is presented that includes the specification of Quality of Services (QoS), along with the required computational resources for service execution in CoCs. Additionally, this paper presents an infrastructure that enables the execution of such workflows in CoCs. Finally, a case study using SHARCNET as a CoC is discussed in order to evaluate the proposed approach.