Energy-saving Task Scheduling Algorithm From Three Decision Perspectives

As a new type of computing mode,cloud computing has become the attention of various fields by virtualization technology and cloud distributed approach.With the continuous expansion of cloud computing clusters,there is a lot of waste of energy in cloud data centers,which is a vital problem.Among the many technologies of cloud computing,task scheduling can effectively reduce energy consumption.Many papers of cloud task scheduling researches mainly focus on completion time or load balancing,and lack systematic consideration of task goals.Aiming at the problem of energy consumption optimization,this paper takes cloud task scheduling technology as the basic starting point for solving the problem of energy saving optimization.It can save energy consumption and improve resource utilization by using task scheduling technology,which has important theoretical significance and practical value.This paper focuses on the task scheduling technology in the cloud environment,analyzes and compares the key issues of the cloud task scheduling technology.In view of the cloud task scheduling method to solve the question of dividing the task request resources,the cloud task request resources are divided into three types: CPU type,memory type,and hybrid type through three-way clustering methods.The theory of greedy algorithm is introduced into cloud task scheduling technology,and a series of researches of energy-saving task scheduling of cloud task request resources from the perspective of three-way decision are carried out.(1)Aiming at the problem of low host resource utilization.From the perspective of cloud task division,we propose the task energy-saving scheduling model based on threeway clustering.The method takes tasks as the basic element of resource consumption,it also considers the heterogeneous characteristics of the host and three-way clustering are introduced into the cloud task division.By analyzing the requested by resources of cloud tasks firstly,and then add a hybrid task division method to divide the types of cloud tasks and combining the basic idea of three-way clustering,the TCKTW algorithm is proposed to divide cloud tasks.The algorithm divides the trivial domains of each cluster into hybrid tasks and divides cloud tasks into three types.It places them on appropriate host resources with scheduling strategies,and propose a research on task energy-saving scheduling based on three-way clustering.Finally,simulation experiments show that there is a certain degree of reduction in energy consumption,by comparing to k-means,k-modes,and k-medoids clustering algorithms,(2)The traditional single-objective task scheduling research only focuses on the total completion time of task scheduling or load balancing.In this paper,considering the dualobjective problems and the greedy algorithm is introduced into task scheduling,and then the related research on task scheduling is proposed,which is the greedy scheduling strategy based on TCKTW.Firstly,analyzing the existing task scheduling problems and combining the basic ideas of the greedy algorithm to establish a cloud computing task scheduling model based on greedy.Then,the cloud tasks are divided into CPU type,memory type,and hybrid type based on the TCKTW algorithm.With the goal of minimum execution time and satisfaction as the goal,finding the optimal solution and constructing a kind of greedy scheduling model based on that algorithm.Finally,a simulation experiment is carried out on the model proposed in this paper.The experimental results show that it is in line with the existing comparing the two algorithms Min-Max and MIN-Max-Min,the model proposed in this paper can effectively improve resource utilization,shortening the total completion time of tasks and save energy.