| Currently the resources of the cloud computing environment are mainly the virtual machines. Its core idea is that virtual resource pool is formed by virtualization of hardware resources in data center. Then through resource scheduling management those resources were provideed to users transparently with “on demand,pay as you go” style. With the increasing expansion of the number of users and the scale of the data center, the major challenge which the cloud computing faces is how to deploy such virtual machine resources rapidly, efficiently and dynamically. The efficiency of virtual machine resource scheduling will directly affect the overall performance of cloud platform. Thus, to meet the user Quality of Service(QoS) requirements and the shortest task total execution time and not to violate the service level agreement(SLA), how to improve data center resource utilization rate is a major problem for cloud environment resource scheduling management.Resource scheduling in cloud environment is a process that virtual machine resources in virtual resource pool are allocated to the tasks submitted by users. When a large number of users submit tasks, we must develop a rational and effective resource scheduling strategy to make the best of resources. This paper mainly studies virtual machine resource scheduling strategies in cloud environment and the major research works and innovative points of this paper include the following:1. Firstly, we analyze systematically the research background and siganificance of cloud computing, as well as the domestic and foreign development and present research situation and introduce cloud computing and some key techniques for implementation of cloud computing. Then with the careful analysis of the characteristic of virtual machine resources and virtual machine resource scheduling, this paper summarizes the virtual machine resource scheduling model and goal for cloud computing, and advantages and disadvantages of commonly-used resource scheduling algorithm.2. Comparatively analyzing cloud datacenter’s resource schduling strategies of some well-known IT vendors, such as Amazon, IBM and so on. The resource scheduling problem in the environment of the cloud computing is studied. In view of the problem that resources utilization rate is not high and node load imbalance problem in current cloud computing environment, firstly this paper studies a new scheduling strategy on load balance of VM(Virtual Machine) based on genetic algorithm. Acoording to historical data and current state of the system and through genetic algorithm, this strategy can achieve the best load balance and reduce or avoid dynamic migration. Then this paper discusses a virtual machine resource scheduling algorithm based on modified genetic algorithm. This algorithm makes a goal oriented initial population by the virtual list of resources and their updated properties. For making the initial population, it uses two stages evaluation. At first stage, all user tasks will be ordered by a priority method with respect to their influence to each other based on graph topology. In the second stage, the candidate resources will be asigned by combination of features of two methods, Best-Fit and Round Robin to select good candidate resources. Finally, applying crossover and mutation operations to achieve the best solution.3. To verify the feasibility and effectiveness of the algorithm, this paper uses the cloud simulation platform——CloudSim which is developed by Melbourne University to simulate virtual machine resources and user tasks under cloud environment, make simulation to the two algorithms discussed in this paper. Through the analysis and comparison from multiple aspects, the two algorithms which is more suitable for the cloud environment can achieve good load balancing and improve the resource utilization rate has shorter task execution time and higher execution efficiency. Finally, after summarizing the full text, this paper puts forward the future research direction. |
PDF Full Text Request