Research On Complicated Resource Scheduling Based On Energy-aware In Cloud Computing

In recent years, with the development of the IT new technology represented bycloud computing, Internet of things and big data, the users of the technologyinnovation platform are increasing in numbers. The information processingrequirements of large numbers of users lead to the growing scale of the data center oftechnology innovation platform. The inappropriate resource scheduling brings abouthuge amounts of energy consumption in the face of mass information processingrequirements. As a result, the operation cost of technology innovation platform risesand numerous environmental problems appear. Therefore, the high energyconsumption problem in the data center of technology innovation platform is studied.The characteristics of application requirements of technology innovation platformand the factors to high energy consumption of technology innovation platform datacenter are analyzed. In the light of the theory of Green Cloud and the practices ofbuilding green cloud data center, the green cloud architecture of technologyinnovation platform is put forward. And, the technology innovation platform datacenter is sure to be low-carbon and energy-saving at the system level.In the data center of technology innovation platform, appropriate virtual machineplacement policy can reduce energy consumption and improve the Quality of Service(QoS). However, the randomness and abruptness of user arrivals and the complexityand multidimensionality of resource requirements increase the difficulty of virtualmachine placement. Thereby, the energy-aware optimal scheduling problem withcomplicated resources requirement is modeled. Multi-dimensional space partitionmodel is presented to guide the placement of virtual machines, so as to decrease thesizes of resource fragments and satisfy QoS. Based on this model, an energy efficientonline virtual machine placement algorithm OEEVMP is further proposed. At last,OEEVMP is evaluated with online first fit algorithm OFF via extensive simulationsand experiments. Experimental results demonstrate that OEEVMP can reduce the sizes of resource fragments, decrease the number of running physical machines, andthus lower the energy consumption. Moreover, experiments on real trace of Googleverify the feasibility and validity of the proposed algorithm.The theories and techniques of Green Cloud, vector bin-packing and ArtificialIntelligence are applied to solve the high energy consumption problem in technologyinnovation platform. In the system aspect, the green cloud architecture of technologyinnovation platform is brought up. In the technical aspect, multi-dimensional spacepartition model is used to solve the energy-aware optimal scheduling problem withcomplicated resources requirement. OEEVMP is proposed to solve online virtualmachine placement problem. The proposed model and its algorithm not only satisfyoptimal energy-efficient, but also fulfill timely response to users’ requirements. And,the goal of energy-saving and the guarantee of QoS are achieved in technologyinnovation platform data center. Through a combination of the model and the greencloud architecture, the optimization of energy consumption can be reached in bothsystematical and technical aspect. Eventually, the technology innovation platform isbound to be green, low-carbon, energy-saving and environmentally friendly.