Energy-efficient load balancing for divisible tasks on heterogeneous clusters

With the growing power consumption of heterogeneous clusters, energy-aware resource management has been a hot research topic recently. Among various candidate techniques, load balancing is one of the software schemes for energy optimization through mapping tasks to servers properly, especially for divisible tasks like web and database queries. However, most practical load balancing schemes are lack of energy awareness, while recent energy-aware schemes usually have high time complexities. In this article, we present an energy-efficient load balancing (EELB) scheme that consists of two-phase heuristics, including server provisioning and load balancing. Based on prior knowledge of power models of servers in the cluster, the proposed scheme is able to make wise decisions on selecting power-on servers as well as scheduling tasks to servers. To validate EELB, we conduct extensive experiments on both simulated and real-world testbeds. Experimental results reveal that the proposed method outperforms competitors in terms of energy saving, and is more applicable to heterogeneous clusters.