The Research Of Dynamic Resource Management Key Technologies In Cloud Computing

In the background of the rapid development of the Internet, cloud computing hasbecome a hot topic today as a new mode of information processing. Supported withvirtualization technology, cloud computing provides a dynamic and scalable serviceaccording to the needs of users. With the popularity of internet and its enhancedtechnology, the total amount of resources surge and its dynamic resource managementand allocation of scalable features, which led to the management and scheduling ofresources become critical impact on cloud computing performance. As so far, therestill exist many problems to be solved in resource scheduling and management ofcloud computing, with the background of cloud computing technology, this paperfocuses on studying the cloud environment scheduling and optimizing allocation ofresources from the load balancing, genetic algorithms and market economy models tosearch cloud environment and efficient resource management programs, the majorresearch work are as follows:1).In view of the weighted least connection scheduling algorithm of fixed weightcausing the load imbalance problem, this paper put forwards the probability predictionscheme based on dynamic feedback, dynamic adjustment algorithm of weights, andthrough the probability forecast to improve the accuracy of weight in a dynamicenvironment.2).Proposed a genetic algorithm to provide solutions based resource, whichmainly using the tree structure and three-dimensional segmentation of chromosomesto encode and decode using the spanning tree algorithm to generate the initialpopulation of virtual machines, Designs the appropriate fitness function evaluationprogram and the thesis. Based with the structure change of chromosome, designsselection operator, cross operator and mutation operator. Experimental results showthat the algorithm compared with the traditional algorithm, the convergence timeshows faster while less times of virtual machine migration.3). Focused with the limit of existing research uneven cloud marketing, this paper proposes resource bidding strategy particle swarm double auction by analyzing thecloud data center resources available bid scheduling problem. Combined with theinertia weight, field topology, Particle Swarm Optimization (PSO) learning factorsand reproductive processes, this paper optimizes particle swarm algorithm to make itmore adaptive to the cloud environment.The simulation experiment verifies algorithm proposed in this paper feasible andeffective, shows a high superiority compared with others.