A Game-Theoretic Analysis of Energy Efficiency and Performance for Cloud Computing in Communication Networks

Abstract

Continuing growth in cloud-based services and global Internet protocol traffic necessitates performance improvements in energy consumption, network delay, and service availability. Data centers providing cloud services and transport networks have often multiple stakeholders, which makes it difficult to implement centralized traffic management. This paper adopts a game-theoretic approach to data traffic management to obtain a distributed and energy-efficient solution, where each edge router is acting as a strategic player. A multiobjective optimization problem with a priori user-specific preferences is formulated for each player, and a distributed iterative algorithm is proposed to solve the game. The existence of Nash equilibrium of the proposed game is shown, followed by the theoretical convergence analysis of the iterative algorithm. The efficiency loss between the strategic game and the corresponding global optimization method is analyzed to quantify the impact of selfish behavior on the overall system performance. Simulation results show notable challenges for operators to plan, design, and operate a multimedia content network to optimize energy consumption, network delay, and load balance over a diurnal cycle.