Research On VM Resource Management Based On Energy Consumption And SLA In Cloud Environment

The transaction between users and the Cloud service agent is completed throughthe International. In cloud,‘everything is service’, every confrontation is a trade.Based on grid computing, parallel computing, etc. and taken virtual technology asthe cornerstone, Cloud computing has developed into another innovative generationtechnology product of the21st century comply with times. Cloud computing makesfull use of the Internet platform, to provide the virtual computing resources, storageresources and cyber source according to the user needs to the user, and to perfectlycomplete the transaction, SLA protocol between the users before the transactionmust be signed. The increasingly mature business model and technology architectureof Cloud computing makes Cloud computing has become one of the hottesttechnologies, which has brought a large number of users resorting to cloudcomputing, and to meet these user’s requirements, more IT devices, refrigerationequipments need to be set, and more mass data storage must be put on the cloud, inaddition, the performance requirements of Cloud services from these users arebecoming increasingly specific and stringent. Therefore, researches on the strategiesof energy saving and SLA constraints in the process of cloud resource provision areimportant theoretical and practical significanceThis article has analyzed and studied the provision process and allocationpolicy for virtual machine resources in cloud environment in depth and width,analyze some existing management strategies of virtual machine resources and thenfigure out their advantages and shortcomings. Here the work mainly focus on twoaspects, namely virtual machine resource provision based on mixed genetic&simulated annealing algorithm, and resource allocation based on evolutionarygame theory.(1) The proportion of Energy consumption in the process of resourceprovisioning is relatively high; and the performance requirements from the cloudusers constrains the cloud resource providers cannot optimize resources provisionstrategic by simply saving the energy consumption, these problems are considered inthe proposed strategy in this paper, by using a existing energy consumption model, we build energy function, turning the function into the fitness function whichevaluate the genetic capacity of in mixed genetic and simulation annealing algorithm,and the SLA performance constraints into the resource sharing rate of physicalserver, and set up a shared rate threshold, so the server to achieve optimaloperational performance in order to meet SLA agreements. Results obtained bysimulation of different strategies in CloudSim Simulation showed that the mixedgenetic simulated annealing algorithm is better than genetic, simulated annealing,and is faster, more accurate for the optimal policy values.（2）From the point of the economic view, applying the thought of game theoryto the cloud resource allocation optimization, which can give full play to theadvantages in the process of resource management,Computing resource allocationscheme is proposed based on evolutionary game theory clouds, introducing geneticand mutation factor that can more accurately describe the evolution and impactperformance and variation played by genetic factors, in addition to the allocation ofresources in the process of energy and quality of service into account, the decisionresource pricing programs, and through the metering utility function assessment, andfinally realize the value of the utility function optimal resource provisioning policy.