Energy-aware scheduling of service requests in multi-tier environment: Analysis and evaluation

Abstract

In a multi-tier system having heterogeneous servers, service requests are usually allocated to available server nodes based on scheduling policies to enforce performance. Such an approach leads to underutilized servers in many modern IT service centers during average low workload periods. Looking to increase overall servers efficiency and having slight performance degradation, in this paper we propose a new power-aware routing algorithm for dynamically scheduling services requests that takes into account the tradeoff between energy consumption and performance requirements. Experiments performed in a simulated environment have demonstrated the energy savings of our algorithm against well-known routing strategies such as shortest response time and least server utilization considering three different workload scenarios.