VM Selection Process Management for Live Migration in Cloud Data Centers

With immense success and fast growth within the past few years, cloud computing has been established as the dominant computing paradigm in information technology (IT) industry, wherein it utilizes dissipated resource benefits and supports resource sharing and time access flexibility. The proliferation of cloud computing has resulted in the establishment of large-scale data centers across the world, consisting of hundreds of thousands, even millions of servers. The emerging cloud computing paradigm provides administrators and IT organizations with considerable freedom to dynamically migrate virtualized computing services among physical servers in cloud data centers. Normally, these data centers incur very high investment and operating costs for the computing and network devices as well as for the energy consumption. Virtualization and virtual machine (VM) migration offers significant benefits such as load balancing, server consolidation, online maintenance and proactive fault tolerance along data centers. VM migration relies on how to determine the trigger condition of VM migration, select the target virtual machine, and choose the destination node. As a result, dynamic VM migration in the scope of resource management is becoming a crucial issue to emphasize on optimal resource utilization, maximum throughput, minimum response time, enhancing scalability, avoiding over-provisioning of resources and prevention of overload to make cloud computing successful. Intelligent host underload/overload detection, VM selection, and VM placement are the primary means to address VM migration issue. Therefore, these three problems are considered to be the most common tasks in VM migration. This thesis presents novel techniques, models, and algorithms, for distributed dynamic consolidation of virtual machines in cloud data centers. The goal is to improve the utilization of computing resources and reduce energy consumption under workload independent quality of service constraints. The proposed approaches are distributed and efficient in managing the energy-performance trade-off