The Research For Resoure Scheduling Strategy And Simulation Analysis In Cloud Computing

The concept of cloud computing is proposed by Google, the major companies in the industry have invested in research for many years and introduced a lot of new products. Government over the world issued relevant policies to support domestic enterprises. Cloud computing technology is evolved from parallel computing, grid computing, utility computing and virtualized computing. It is a mode of on-demand services. As a development, Cloud computing has gradually become the hot spot of research institutions and commercial organizations. So how to allocate resources reasonably to cater requirements from both sides attracts wide attentions. A good resource scheduling needs to balance the interests of all parties at the same time, resource scheduling goals are not the same. There are many research objects, such as reducing the energy consumption in the data center, improving service quality, improving resource utilization based on econometric models, load balancing, etc.The key technologies of resource scheduling scheme include: scheduling strategy, optimization goals, scheduling algorithms, data center infrastructure. The scheduling policy is formulated by data center managers,it is the top-level strategy for scheduling management. Optimized goals and the need of data center are related,scheduling will determine whether the scheduling quality is good or bad. Optimized goals include speed of response, quality of service, and total cost control. Rational allocation of resources need dynamic allocation of cloud computing data center’s physical and virtual shared resource. Not only considering the performance of the dynamic system, but also considering the cost of the data center, which cloud computing data center resources to efficiently manage scheduling algorithm cater different business need. The key to establish a cloud computing data center platform is how to dynamically manage and allocate resources and improve resource utilization.Map/Reduce is a programming model for cloud computing by Google company, and has been widely used in the cloud computing research area. Map/Reduce is composed of two parts:The first part is to divide the original task into multiple sub-tasks; the second part is to assign to virtual resources of sub-tasks, this thesis is based on the second part of the Map/Reduce thought and do the work in two areas:(1)Resource scheduling problem, this paper studies how the various sub-tasks are assigned to virtual resources, in order to improve the overall completion time in the task. Cloud environment resources ared modeled as task scheduling genetic optimization and ant colony optimization algorithm combination optimization problems by the improved algorithm, the improved genetic optimization algorithm to obtain the initial solution, then the initial solution substituting improved ant colony algorithm to obtain the optimal solution. Combination of genetic optimization algorithm global optimization ability of ant colony optimization algorithm of local optimization capability. Cloudsim simulation results show the superiority of the algorithm, experiments show that the algorithm converge faster optimization capability, and it is an effective task scheduling algorithm.(2)Data center energy consumption problem, the paper proposes a virtual machine migration strategy considering multiple factors, by choosing the minimum migration cost and the minimum energy consumption of network transmission cost to reduce the energy consumption of the data center. The experiment results can tell that this strategy can effectively reduce the total cost of network traffic and migration. This strategy can be a part of the contribution to save the cost of the data center.