With the emergence of cloud computing which can deliver on-demand high-performance computing resources over the Internet, cloud workflow systems offer a competitive software solution for managing large numbers of business processes. In this paper, we first analyse the basic system requirements through a motivating example, and then, the general design of a cloud workflow system is proposed with the focus on its system architecture, functionalities and QoS (quality of service) management. Afterwards, the system implementation of a peer-to-peer based prototype cloud workflow system is demonstrated to verify our design. Finally, experimental results show that with the dynamic resource provisioning, conventional violation handling strategies such as workflow local rescheduling can ensure the on-time completion of large numbers of business processes in a more cost-effective way.
Funding
Management of Large-Scale Models
Directorate for Computer & Information Science & Engineering
Conferences in Research and Practice in Information Technology Series Volume 127: 10th Australasian Symposium on Parallel and Distributed Computing, AusPDC 2012
Conference name
10th Australasian Symposium on Parallel and Distributed Computing, AusPDC 2012