OCPA: An Algorithm for Fast and Effective Virtual Machine Placement and Assignment in Large Scale Cloud Environments

Cloud Computing is increasingly being deployed as a fast and economic solution to various requests that require instantiating computing resources to serve customers' need. Despite the distinctions among the three commonly accepted service models (i.e., IaaS, PaaS, and SaaS), any large scale cloud computing deployment requires the instantiation of multiple VMs (Virtual Machines) in a coordinated manner. These VMs may further need to be placed in multiple geographically distributed data centers, and users that are closer to a VM enjoy the benefits of smaller response time, higher bandwidth, hence better performance. Thus, how to appropriately place VMs and assign them to cloud users can have significant impact on the latter's performance. In this work, we consider the problems of VM placement and assignment. Based on a set of unique design principles, we propose an effective and efficient solution which is referred to as OCPA (Opportunity Cost based VM Placement and Assignment). OCPA can achieve much better performance and the running time complexity is kept linear.