Green Task Scheduling Algorithms with Speeds Optimization on Heterogeneous Cloud Servers

Currently, a large number of cloud computing servers waste a tremendous amount of energy and emit a considerable amount of carbon dioxide. Thus, it is necessary to significantly reduce pollution and substantially lower energy usage. This paper seeks to implement six innovative green task scheduling algorithms that have two main steps: assigning as many tasks as possible to a cloud server with lowest energy, and setting the same optimal speed for all tasks assigned to each cloud server. A newly proven theorem can determine the optimal speed for all tasks assigned to a computer. These novel green algorithms are developed for heterogeneous cloud servers with adjustable speeds and parameters to effectively reduce energy consumption and finish all tasks before a deadline. Based on sufficient simulations, three green algorithms that allocate a task to a cloud server with minimum energy are more effective than three others that assign a task to a randomly selected cloud server. Sufficient simulation results indicate that the best algorithm among the six algorithms is Shortest Task First for Computer with Minimum Energy algorithm.