Review Article

A Survey of Game-Theoretic Approach for Resource Management in Cloud Computing

Table 3

A comparative analysis of game-theoretic models for resource allocation in cloud architecture [6164].

S. no.Author(s), yearResource allocation strategyResource optimization strategyBenchmark toolsBenefitsLimitations

1Xu and Yu, 2014 [45]A finite extensive form game with perfect informationBackward induction approachHadoop scheduler, first fit algorithm, and Google cluster allocation mechanismPerform better in fair allocation, diminish resource wastage, and attain a better resource utilization rate
2Nezarat and Dastghaibifard, 2015 [46]A repetitive game with incomplete information in a non-cooperative environmentNash equilibriumAchieve Nash equilibrium even with insufficient knowledge of the environment, respond in a shorter period, and provide the lowest violation of the service-level agreement and the most utility to the providerIt is considered a static pricing strategy rather than a dynamic resource allocation method
3Xiao and Tang, 2015 [47]A cooperative gaming modelNon-cooperative gaming modelCommodity market model, double auction model, Vickrey auction model, and batch auction modelReduce the application execution latency and improve the price negotiation efficiency when bundles of resources are negotiated simultaneously
4Yan et al., 2016 [48]Non-cooperative and bargaining game resource allocation algorithmNash equilibrium and Pareto efficiencyHadoop mechanism, Google management mechanism, and general equilibrium algorithmEnsure fair constraints, improve the efficiency of resource allocation, and provide resource competitors with optimal benefits and improved collective benefitsThe slight increase in consumption
5Khansari and Sharifian, 2019 [49]Evolutionary game theoryEvaporation-based water cycle algorithmActive-guided evolutionary game theory and NSGA-IIBe stable, reliable, and optimal; be advantageous for low latency, low bandwidth, and high-security applications, and improve the overall QoSAlthough the optimization problem can be found in a reasonable time, it is not essentially the precise answer to the problem but an acceptable estimation
6Swathy et al., 2020 [50]A Stackelberg game-theoretical modelThe load balancerTraditional random allocation algorithms and flow-shop scheduling algorithmReduce the number of task failures, increase throughput and high resource utilization, decrease ‘makespan” value, and decrease the error rateThe whole load balancing process cannot be achieved by just executing this algorithm that is a long-term process
7Carlucci et al., 2020 [51]A non-cooperation model based on minority game theoryA decision modelThe classic formulation of the minority gameBe adequate in allocating resources and react to an unexpected event with robustnessThe use of homogeneous resources shared in the network
8Zhanga et al., 2020 [52]Integer programming model for the time-varying multidimensional resource allocationWaiting period strategy and dominant-resource-based strategy to improve the social welfare and resource utilizationAddress the problem of online time-varying multidimensional resource allocation and pricing in clouds