Energy Guided and Workload Adaptive Modeling for Live Migration
Bing Wei · Advanced materials research · 2013
Live migration provides desirable benefits in the field of energy saving, which packs service into fewer physical servers while maintaining the performance level. In this paper we present energy guided and workload adaptive modeling for live migration, two models are developed respectively including energy guided migration model and workload adaptive model, the former model selects the best migrating virtual machine (VM) candidate with the minimal energy consumption while the later model chooses the best migrated physical server candidate in terms of both energy and workload characteristics, furthermore, concerning the service quality, workload adaptive model also takes charge of the determination of the live migration moment. The experiments results show that our approach achieves significant energy saving and robust live migration.