QoS-Aware Service Redeployment in Cloud

Service composition is a useful technology to achieve dynamic requests. But in cloud, the cloud provider should guarantee the QoS of the user request as well. Previous work investigated how to select proper available replicas (the similar functional services located on different physical nodes) to effectively achieve the composite services. However, as the requirements for some service grow, even the optimal selection strategy could not satisfy all the QoS requests, if it is only based on the existed replicas. In this case, cloud provider should deploy more replicas to meet the growing requests. Some existing literatures have addressed the service redeployment problem, aiming to optimize the overall or average performance; however, few of them can guarantee QoS of each request. This paper investigates QoS-aware service redeployment problem (SRP), with objective to minimize the redeployment cost. We show that, it is NP-hard to decide whether there exists a feasible solution of SRP. Thus we propose a novel heuristic algorithm SRA, which can find a solution such that most of the requests can be satisfied, while the deployment cost is minimized. Experimental results show that our approach is effective and efficient.