Research On Fairness Allocation Strategy With Efficient Resource Use In Cloud Computing Environment

Recently,with the rapid development of cloud computing, a growing number of users and organizations run their business in the cloud platform. This condition has led the heavier load of the cloud datacenter. The cloud datacenter integrates several kinds of resources to realize uniform distribution and management. The core problem of resource allocation is to solve the problem of multi-resource virtual machine allocation. The cloud datacenter assigns resources rationally to response the request of resource. Rational distribution means good fairness, fitness and resource utilization. However, the traditional allocation algorithm has the problem of low resource utilization and unfair distribution. Regarding the issue above, we proposed an allocation method of improving the resource utilization and fairness in cloudcomputing. The thesis mainly does the following work:(1)We have abstracted and modeled the cloud datacenter and give the detailed formal expression of the datacenter including virtual machine, physical machine,resource utilization and etc to make it easy for us to design resource allocation strategies and analysis experiment.(2) The resource allocation method was proposed based on bin-packing strategy in cloud computing environment. Based on the description and analysis of the whole problem, the definition and constraint condition of the method are given. The method is based on the type matching of the virtual machine and the physical machine, so the article gives a specific classification algorithm and the final objective function of the method. The method first uses Fist Fit Decreasing algorithm to achieve the goal of reasonable resource allocation and balanced resource allocation in the early stage. And then the method uses the Best Fit Decreasing algorithm to solve the problem of maximizing the use of resource in latter part of the allocation. Finally, a simulation experiment is carried out to compare with the Fist Fit Decreasing algorithm and the Best Fit Decreasing algorithm. The experiments show that our method has higher resource utilization.(3)The fairness-optimized resource allocation method was proposed in cloud computing environment. Based on the description and analysis of the problem, the formal description of the main parameters is given. Use of Global Dominant Resource Share and Global Dominant Resources Weight is to solve the problem of user priority confirmation and qualification queue confirmation. And fairness are defined and explained in theory, including establishing an evaluation model for the fairness. The main process and the fairness properties of the method are analyzed and described. Finally, the method is experimented and compared with the algorithm of maximizing multi-resource fairness based on dominant resource, which has better superiority in the allocating number of virtual machines,the utilization rate ofcomputing resources, the utilization rate of memory resources and the whole fairness.