Phase Space Based Energy Consumption Model and Optimization Analysis in Clouds

The Cloud is a highly coupled system with massive nodes, and its energy problem has gradually become the focus of researchers. Existing research lacks an overall depiction of the cloud clusters' energy state. Therefore, the complex cloud system is reconstructed through the phase space to research the cloud energy from an overall level. Firstly, we propose an energy prediction model that considers the server node's context to predict the tasks execution energy on the node. Secondly, an energy phase space of the cloud system is established, including a static energy phase space that represents the current overall energy state of the cloud system and a dynamic energy phase space that represents the energy increment when the Cloud receiving tasks. Finally, a Phase Space-based cloud task scheduling Algorithm (PSA) is designed. The evaluation results show that PSA is superior to other algorithms in the scheduling performance, and the larger the cluster size is, the more obvious its advantages.