Performance Interference Attentive Live VM Migration in Cloud Computing

Abstract

Cloud Computing can be thought as a multi-tenant system where multiple users can host the applications on cloud server. Cloud computing is the conveyance of computing services including servers, storage, databases, networking, software, analytics, and intelligence over the Internet to provide faster revolution, elastic resources, and financial prudence of scale. Users typically pay only for cloud services as they use, serving at lower operating costs, run infrastructure more efficiently and scale as the business needs change. Hosting multiple applications on single physical cloud machine by applying virtualization leads to maximum resource utilization of cloud server component. But when resource demand of certain VMs goes high than the maximum allowable capacity the running VM need be migrated from current physical server to other physical server of the cloud infrastructure. Live migration of VMs is a crucial feature of virtualization, which allows migrating VMs from one physical machine to another physical machine without suspending VMs. A cloud administrator can migrate an instance with application so that the machine can be made available for maintenance. Similarly, to improve manageability, OS instances may be rearranged across machines to relieve the load on overloaded hosts. To perform the live migration of a VM, its runtime state must be transferred from the source to the destination while VM still running. Live migration can lead performance degradation because of migration event this is known as migration interference and the migrated VM on the destination PM can grab the shared resource portion on the destination PM after migration which may lead the less recourse usage among the co-located VMs this down effect is known as co-location Interference. This research project tackles the research challenges that are related to virtual machine selection process but also are related to physical machine selection process by keeping the inference minimization as the centre focus.