Research And Implementation Of Cloud Computing Platform Scheduling Management Technology

With the continuous maturity and development of IaaS cloud platformtechnology, more and more small and medium-sized enterprises and applicationservice providers deploy their applications to cloud platforms, which results incontinuous expanding of the cloud data center, more complexed virtualizationenvironment and rapidly increased consumption of energy. Currently, the problems ofresource monitoring, virtual machine scheduling and energy consumption in cloudplatform have aroused extensive attention in academia. Today, advocating foreconomic, energy-saving and environmental protection, how to protect the user SLAagreement implementation, and meanwhile improve data center resource utilization,reduce the resources investment costs and energy costs, so as to maximize theeconomic benefits? This presented a new challenge to the cloud platform providers.This is also hot spots and difficulties in cloud computing platform schedulingmanagement technology. To solve this problem, the main work and contributions are asfollows:Firstly, this dissertation analysed the parameters framework of the SLA, SLAfeatures and technical evaluation parameters of cloud platform services, and then onthe foundation of the SLA agreement research, proposed SLA-based resourcesmonitoring scheme. This scheme includes a new type of cloud computing environmenthierarchical resource monitoring model and a SLA-based resources monitoringstrategy. The experimental results show that, compared with the traditional scheme, theSLA-based resource monitoring scheme has a better consistency index, so as to betterensure the SLA management and implementation of the cloud platform.Secondly, for the resource scheduling technology of cloud computing platform,this dissertation combined the verified SLA-based resources monitoring scheme, andproposed a SLA-based virtual machine energy-saving scheduling mechanism. Thismechanism includes a SLA-based virtual machine elastic stretch strategy strategy anda SLA-based virtual machine migration scheduling strategy. The experimental results show that, compared with the traditional resource scheduling algorithm, the SLA-basedvirtual machine energy-saving scheduling mechanism has lower SLA violation rate andbetter energy-saving.Finally, this dissertation designs and implements the cloud platform system whichis based on the SLA monitoring and scheduling strategy. Resource monitoring strategyand virtual machine scheduling strategy are applied to the system’s real-timemonitoring, instance management and strategy management module. This dissertationdetailed exposition the design and implementation of the system, proved the scientificand feasibility of the proposed scheme from the theory and practice. It has importantsignificance in the cloud computing platform scheduling management technology.