Research On Three Clustering Algorithms And Their Application In Cloud Task Rapid Scheduling

With the continuous development and maturity of cloud computing industrialization,it has gradually become one of the essential facilities for developing many high-tech industries.As a commercially distributed computing mode,cloud computing provides cluster resources to users in the way of on-demand services.With the expansion of the industrial scale,the energy consumption and resource waste generated in cloud data centers become the focus of cloud resource management technology.As one of the essential energy-saving technologies in cloud computing,task scheduling focuses on allocating tasks submitted by users to appropriate VMS to manage cloud power consumption effectively.Hence,cloud task scheduling technology is an important research topic of cloud computing resource management.Cloud task scheduling technology is related to performance indicators related to economic interests between users and cloud service providers,such as task completion time,system load balance,energy consumption,resource utilization,etc.The traditional clustering method for cloud task scheduling is a practical resource management technology.However,some fuzzy cases exist in the cloud task division,especially in middle and cloud task overlap under the real-time cloud environment,that is,the clustering of cloud tasks is not a two-way classification problem.The improved clustering method based on the three-way decision theory can often effectively solve such cloud task division problems.Based on the primary feature of cloud task and scheduling technology,this paper presents a three-way cluster of cloud tasks and their application under a cloud environment.The main contributions of this paper are as follows.(1)We propose a cloud task granulation method based on three-way clustering(TWC-TG)by integrating the three-way decision and cluster algorithm into the cloud task technology.In this part,we firstly divide tasks into various clusters according to cloud task attributes,task request computation amount,task memory,and network bandwidth,and any cluster consists of two parts: core region tasks and edge region tasks.Meanwhile,these granulation tasks are placed on fitting host resources via some scheduling strategies.Finally,the experiments on the Cloudsimplus platform show that our method can effectively shorten the task completion time and reduce the energy consumption in the data center.(2)We propose a dynamic programming scheduling algorithm based on three-way clusters(TWOCP)for the problem of optimization scheduling strategy of cloud tasks.In this part,the three-way clustering algorithms are combined to divide the cluster according to the attributes of cloud tasks,and dynamic programming is used to schedule tasks in the core region and edge region of each cluster.Compared with greedy and genetic algorithms,dynamic programming algorithms are a global optimal scheduling strategy and cannot easily fall into optimal local conditions.Finally,the simulation results based on the Cloudsimplus experiment show that the algorithm significantly reduces the task completion time and the energy consumption and effectively ensures the availability of the data center.