Two-tier policy-based consolidation control for workload with soft deadline constrain in virtualized data center

Resource virtualization has emerged as a promising approach to solving the under-utilized caused by over-provisioning and time-varying resources requests in data center, and rational consolidating workload in a virtualized infrastructure has recently gained considerable attention. A two-tier policy-based consolidation controlling scheme for workload with soft deadline constrain in virtualized data center is proposed. It allows the resources requests to be moderately beyond the resources capacities while allocating both virtual and physical resources on-demand, for the purpose of achieving higher server utilization and energy efficiency without taking the deadline penalty in a best-effort manner. A test-bed with a virtual data center using Xen virtual machines is built based on CloudSim. The experimental analysis indicates that the desired system goals can be leveraged to improve the overall utility.