Abstract:

In view of the characteristics of data intensive operations, we proposed an evaluation model based on CPU and graphics processing unit (GPU) as two influencing factors to calculate the energy consumption of the nodes. The model based on the selection of the original virtual machine (VM) energy saving algorithm in the virtual machine selection of energyefficiency advantages, it further used heuristic idea to improve the bee colony optimization algorithm, and sought the best optimal solution of VM integration. The experimental results in the CloudSim 3.0 cloud computing simulator show that the heuristic backward bee colony algorithm can effectively reduce the number of VM migration on the premise of guaranteeing the quality of service, and then reduce the energy consumption of data center (25%—30%).

Key words: data center, virtual machine migration, virtual machine integration, artificial bee colony algorithm