查看文章

Cloud computing has emerged as a modern platform in the era of utility computing by enabling elasticity and on-demand provisioning of computing resources. The proliferation of cloud computing has resulted in the foundation of small to large scale cloud data centers (CDCs) that are able to deliver guaranteed services and ensure the availability of resources. However, the CDC consumes a huge amount of electricity to keep servers running and for cooling. Thus, there is a need to propose a novel framework that ensures efficient utilization of cloud resources while avoiding over-and under-provisioning of resources. Over-provisioning of resources arises when the application is allocated more than the required resources, and under-provisioning occurs when the application gets fewer resources. Thus, over-provisioning leads to under-utilization of resources and needless energy consumption. While under-provisioning leads to overutilization of resources, service level agreement (SLA) violation, and performance degradation. For such a problem, server consolidation, or dynamic virtual machine (VM) consolidation, is an effective solution because it handles and avoids under-and overutilization problems. Also, server consolidation solves the problem of load imbalance by avoiding VM placement on underutilised and overutilized hosts. Any solutions developed for server consolidation have to provide answers to three research questions:(i) When to migrate VMs?(ii) Which VM (s) to migrate?(iii) Where to migrate the VM (s)? In particular, in this paper, we propose a framework for server consolidation that answers all three questions about server …