Virtual machine placement in service-oriented computing environments

Service-oriented computing (SoC) is a growing environment due to its ability to develop distributed applications at low-cost and rapid composition. The building block of the SoC environments is the service. A service should be hosted by a virtual machine (VM), and a VM is hosted by a physical machine (PM). Subsequently, a number of services cooperate to perform certain tasks of an application. These cooperating services need to communicate to perform the tasks of the distributed application. Thus, the proper placement of VMs that host cooperating services is a crucial factor that affects the performance of the distributed application. In this context, we proposed addressing the problem of VM placement in SoC environments as a bi-objective optimization problem. This optimization problem consists of two objectives, namely (1) reducing the number of PMs hosting the services and (2) reducing the communication of the cooperating services forming the distributed application. The proposed method utilized the well-known method of NSGA-II due to its remarkable performance in addressing similar placement problems. Then, the proposed method is thoroughly evaluated on several datasets and compared against two heuristics and one meta-heuristics method. The obtained results show that the proposed method outperformed the methods of comparison by a considerable margin.