Optimization Of Service Resource Allocation Mechanism Based On Game Theory

Facing the urgent need of cloud computing system cyberspace and improving service quality,and the ever-increasing of end user scale,the optimal allocation issue of service resources has attracted lots of attention.Nowadays,the pricing problem,the energy-saving problem,and the utilization problem of computing service resources,as well as the limited problem in the amount of security service resources have become a series of challenges of optimal resource allocation.However,the existing work of the service resource allocation mechanisms needs to be improved under the background of real-time task uncertainty,the conflict between service quality and energy consumption,target uncertainty of malicious attack strategy and the large-scale of potential attack targets.In order to solve the above-mentioned challenges,mainly based on game theory,this paper proposes a number of computing service resource allocation mechanisms to resolve the conflict between service quality and energy consumption under real-time task uncertainty,and in terms of the limited number of security service resources,two kinds of security service resource allocation mechanisms have been proposed to efficient detect out the malicious attack strategy targets and protect the security of the large-scale cyberspace.1.Specific to the reasonable pricing problem of computing service resources,this paper proposes a pricing allocation mechanism of computing service resources.The cloud service provider,fog agents and users are modeled as a two-layer Stackelberg game,and the optimal bids for one unit and prices for one unit of fog agents,the optimal resource allocation of the cloud service provider,and the optimal task allocations of users can be obtained by calculating the game equilibrium.This mechanism brings the more resource profits for cloud service provider on the premise of guaranteeing the service qualities of users.2.Specific to the energy-saving problem of computing service resources,considering the uncertainty of real-time task and the energy consumption loss of servers restart,this paper proposes an energy-saving dynamic decision-making mechanism of computing service resources.By adjusting the state of servers in real time,the service quality can be guaranteed as well as the system energy consumption can be reduced.3.Specific to the utilization rate of computing service resources,considering the uncertainty of real-time task and the asymmetry of resources and tasks,this paper proposes a matching mechanism between computing service resources and dynamic tasks.Effective matching and task splitting can improve the utilization rate of resources and the submission rate of tasks on the premise of guaranteeing the service quality,as results,the energy consumption of the whole system can be reduced and the resource profits can be enhanced efficiently.4.Specific to the limited number of security service resources,considering the target uncertainty of malicious attack strategy and the large-scale of potential attack targets,this paper proposes two security service resource allocation mechanisms by taking denial of service(DoS)attack and man-in-the-middle(MITM)attack as examples,respectively.Specifically,the dynamic detection and defense mechanism of security service resources based on port hopping and singular linear space over finite fields can efficiently detect out the specific attack targets.The security of this mechanism can be completed by theoretical analysis as well as simulation experiment.The multi-objective allocation mechanism of security service resources based on artificial noise packet and heuristic algorithm can efficiently and quickly solve the specific optimal or sub-optimal defense strategy in the large-scale cyberspace system.Both of the above-mentioned mechanisms can achieve considerable protection results with a limited number of security service resources.