Performance Evaluation of Energy-Aware Best Fit Decreasing Algorithms for Cloud Environments

Cloud computing is emerging computational paradigm that provide resources to perform complex tasks. Large datacenters are used to facilitate the incoming tasks by providing resources, such as CPU, memory, storage, and network bandwidth. Datacenters offers hosting and processing of complex tasks and services, where servers and cooling systems consume huge amount of energy. Excessive amount of energy consumption results in large power bills and Green House gases (GHG) emissions. Substantial amount of energy can be saved by powering down servers that are idle. Various authors have come up with energy efficient solutions that try to minimize overall energy consumption. One set of energy-efficient solutions is based on best fit decreasing (BFD) algorithm. In this paper, we evaluate the performance of existing energy efficient BFD algorithms based on various workloads and migration techniques. Moreover, considering the significance of Service Level Agreement (SLA), we introduce SLA-awareness in traditional BFD algorithm to minimize the SLA violation. We present the analysis and observations for each of the considered techniques based on total energy consumption, average SLA violations, and SLA performance degradation due to migration.