The Research Of Resource Scheduling Key Technology In Cloud Computing

Currently, Cloud computing is an emerging computing technology with great potential, which uses large scalable hardware and virtualized resources to provide users with dynamic application services. In order to maximize the utilization of cloud resources and tap the great potential of cloud computing, mining an efficient resource scheduling strategy is our urgent affairs. Resource scheduling strategy is responsible for selecting the appropriate resources from the heterogeneous cloud resource pool to perform the matching task requests. In order to find the optimal scheduling scheme, resource scheduling strategy should be formed in full consideration of the actual scene of cloud computing.Through learning resource scheduling technologies of the cloud computing, deeply analyze the existing resource scheduling algorithms. Then mainly optimized for the cloud resource scheduling strategy under two scenarios, they are designed to improve the quality of services while improving system performance. The main work is as follows:Firstly, there are a number of light weight tasks exist in the applications of cloud environment.If the fine-grained task requirements are allocated to high performance computing resource nodes will increase the overall system latency and turnaround time. A lot of fine-grained tasks will spend a lot of time on task scheduling and transmission. In addition, fine-grained tasks assigned to high performance computing resource nodes will significantly reduce resource utilization. For this cloud scenario, the resource scheduling strategy is based on the dynamic resource priority scheduling policy. And integrate these fine-grained tasks into groups when execute the task. Meanwhile, we consider the status of bandwidth resources to optimize the scheduling strategy.Secondly, the performance of current cloud resource scheduling strategy on resource load balancing resource is not ideal. Most of the existing scheduling algorithms did not consider the differences in user group, resulting VIP users cannot get better services. In order to solve the above cloud system bottleneck, the study is based on the classical Min-Min scheduling algorithm. The low complexity makes the algorithm easy to implement. But the defects of classic Min-Min scheduling algorithm is resource load imbalance. In view of this phenomenon, optimizing the classic Min-Min scheduling algorithm, so the load can distributed evenly and make sure that the service quality of VIP-class user can be guaranteed.Finally, the paper summarizes the contents of the study. Based on the resource scheduling study mentioned before, prospecting a future research direction in this area as the end of research.