Research On Task Scheduling Strategy In Cloud Computing Environment

Physical resource is transformed to virtual by using virtualization technology. These virtual resource isdynamic scalable. Businesses can access cloud resource according to their needs. With cloud computing,users can access a large number of computing and storage resources and get powerful computing abilitywithout considering their specific location and configuration. Each cloud computing system has its ownresource allocation and task scheduling strategy. A good task scheduling strategy can comprehensivelyimprove the operating efficiency of the entire cloud computing system. Therefore, the study has importanttheoretical and practical significance for task scheduling algorithm under cloud computing environment. Inthis paper, some problems that exist in the cloud computing task scheduling algorithm are researched.These problems are resource load balancing, minimum completion time of the task to be executed, taskscheduling algorithm and non-considering the user’s multidimensional QoS requirement. The main contentsare as follows:First, this paper presents a LB-ECT algorithm based on load balancing and minimum completion time.In traditional task scheduling algorithm, it only considers the safety of the task or response time. Inheterogeneous distributed cloud resource pool, the performance of each resource is different. So the systemwhich doesn’t balance, will seriously affect the overall performance of the system in the cloud computingenvironment. If we don’t consider the task execution resources the minimum completion time, the goodperformance of the machine can not be fully utilized. However, many of the existing cloud computing taskscheduling algorithm just consider load balancing or the minimum completion time factors. Thecombination of the two will meet user requirements on overall system performance and time. So, this paperpresents a LB-ECT algorithm based on load balancing and minimum completion time. The algorithmconsiders two factors that is both the resource load and the smallest complete the task in the resources. Itcan get a best match between task and resource by finding an optimal balance point. Simulation resultsshow that the algorithm can improve the success rate of task execution, the utilization of system resourcesand significantly shorten the total completion time of the task.Secondly, this paper presents a cloud computing task scheduling algorithm. The strategy bases on multi-objective particle swarm algorithm with multiple QoS constraints. Currently, Task schedulingalgorithm in cloud computing is mostly centered “machine”. There is no truly “user “as the center. Theproposed QoS-based task scheduling considers the task of user’s needs. But it simply considers aone-dimensional QoS requirement, such as a single time or cost. It can’t meet user’s multi-QoSrequirement. So, this paper presents a cloud computing task scheduling algorithm. The strategy bases onmulti-objective particle swarm algorithm with multiple QoS constraints. The algorithm considers the rate ofthe success, the least cost and optimal make span and so on. While optimizing multiple QoS parameters,this algorithm can find the Task scheduling program that best meets the user’s needs by using the principleof intelligent optimization of multi-objective particle swarm algorithm. Simulation results show that thealgorithm can improve user’s satisfaction, save cost, and improve the execution of the task success rate.