Capability-Aware Energy-Efficient Virtual Machine Scheduling in Heterogeneous Datacenters

Minimizing energy consumption in datacenters requires virtual machine (VM) scheduling methods that effectively deal with system heterogeneities. Existing heterogeneity-aware VM scheduling mechanisms focus on capacity-level heterogeneities while capability-level heterogeneities are largely ignored. In this paper, we propose an integer programming formulation of the VM scheduling model that considers capability-level heterogeneities. Given the NP-hardness of the model, we present a heuristic-based capability-aware VM scheduling algorithm to obtain energy-efficient VM scheduling solutions. Simulation results show that the proposed heuristic algorithm generates near optimal scheduling solutions while significantly reducing computation times.