Research On PDTs Scheduling In Cloud Computing Environment

Cloud computing is the most widely used distributed application system currently. It spreads its computing ability over a resource pool of a large number of servers, which make users can be able to access to computer power, storage space and communication services as required. There are some problems about the current researches on the tasks scheduling in the cloud environment, such as the single task type, poor performance of the scheduling algorithm and not taking into account of the cloud QoS(Quality of Service) at the same time, so the key point of the field of cloud computing researches is how to make the resources distributed more reasonable, and makes the tasks scheduling more efficient so that the tasks users submitted can complete in a short period of time with lower costs and the system can also keep load balancing.On the basis of the analyzing the targets, model, process and characteristic of tasks scheduling, this paper emphatically has done the following works:(1) The article describes the definition, service types, architecture and the features of the cloud computing. And also analyzes the model, objectives and the QoS (Quality of Service) requirements of tasks scheduling in the cloud computing environment and outlines the process and characteristics of cloud computing tasks scheduling.(2) A new tasks scheduling model PDTs (Partial Dependent Tasks Scheduling) based on QoS is proposed on the basis of analyzing the current tasks scheduling model on the virtual machine including independent tasks model and the workflow tasks model. The polymorphic cloud tasks have been analyzed and processed innovatively. And the new model mainly considers the requirements of the users on the deadline of the tasks and minimizing the cost.(3) An improved ant colony algorithm is proposed to solve the PDTs tasks scheduling model on the basis of studying the ant colony algorithm. The algorithm divides the ant colony into several sub ants and choose a better update strategy for the pheromone to get the optimal results by local optimization.(4) An improved particle swarm algorithm is proposed to solve the PDTs tasks scheduling model on the basis of studying the ant colony algorithm. The algorithm also divides the particle swarm into several subgroups and gets the local optimum result and the optimal result by adjusting the velocity and displacement of the particles. (5) On the basis of in-depth study of the CloudSim simulation platform, we do some experiments for the improved algorithms. And the experimental results prove that the convergence and optimum ability of the two improved algorithms are better when compared with the existing algorithms in the CloudSim. In addition, we also compare the performance of the two improved algorithm and the results show the PSO (particle swarm algorithm) is better than the ACO (ant colony algorithm).