Energy optimized modeling for live migration in virtual data center

Live migration has been a significant tool to efficiently pack service into fewer physical servers while maintaining the service quality in data center, and thus provides considerable energy savings by reducing the amount of physical servers. Existed work mostly concentrates on the VM live migration mechanism to satisfy required performance level, however few investigates the energy-guided live migration. We propose a comprehensive approach to optimizing energy consumption by determining the best candidate of migrating virtual machine (VM) and the best candidate of destination physical machine (PM) for new increased workload and migrating VM in an energy cost efficient way, two models are established respectively to settle these two problems: energy aware migration model and load dispatching model. Furthermore, the co-migration that several VMs request live migration simultaneously is discussed and a corresponding heuristic algorithm is developed and evaluated, this paper we believe is the first kind of work to deal with the co-migration decision making in terms of both performance and energy saving. We implement this approach based on Xen 3.4 platform and conduct several experiments which demonstrate that our approach can achieve significant energy saving.