Energy-Aware Cluster Reconfiguration Algorithm for the Big Data Analytics Platform Spark

The development of Cloud computing and data analytics technologies has made it possible to process big data faster. Distributed computing schemes, for instance, can help to reduce the time required for data analysis and thus enhance its efficiency. However, fewer researchers have paid attention to the problem of the high-energy consumption of the cluster, placing a heavy burden on the environment, especially when the number of nodes is extremely large. As a consequence, the principle of sustainable development is violated. Considering this problem, this paper proposes an approach that can be applied to remove less-efficient nodes or to migrate over-utilized nodes of the cluster so as to adjust the load of the cluster properly and thereby achieve the goal of energy conservation. Furthermore, in order to testify the performance of the proposed methodology, we present the simulation results implemented by using CloudSim