Power-efficient resource-guaranteed VM placement and routing for time-aware data center applications

Abstract

Power efficiency and performance guarantees have become major concerns of data center cloud providers as they significantly affect providers’ economic benefits. Providing guaranteed resources necessitates developing a user-friendly and concise request model which accurately abstracts the required server and network resources for a tenant application. We propose a time-aware tenant application (TTA) request model which enables a tenant to express an application request by specifying its resource requirement graph (server resources for VMs and bandwidth) associated with its estimated required time duration. We investigate the power-efficient resource-guaranteed virtual machine (VM)-placement and routing problem for dynamically arriving TTA requests. The problem requires provisioning of the specified resources in a data center for the required time duration of requests by selecting an appropriate set of servers for VM placement and routes for their communication, so as to maximize the number of accepted requests while consuming as low power as possible. We develop a mixed integer linear programming optimization problem formulation based on the multi-component utilization-based power model. Since this problem which is a combination of routing and VM-placement problem, is computationally prohibitive, we develop two algorithms which select servers and routes based on: (1) our proposed goodness function and pre-computed candidate paths, and (2) minimum power cost paths, respectively. We demonstrate the effectiveness of the proposed algorithms in terms of power saving and acceptance ratio through simulation results.