This paper tackles the challenge and study an optimization problem to maximize the cloud gaming provider’s total profit while achieving just-good-enough QoE. We conduct measurement studies to derive the QoE and performance models. We formulate and optimally solve the problem. We conduct extensive measurement studies using an open-source cloud gaming platform, GamingAny-where (GA) on two VM implementations to derive the game-dependent parameters for QoE and performance models. We formulate and propose two algorithms for the provider-centric VM placement problem. We extend the provider-centric VM placement problem into a gamer-centric problem for closed cloud gaming services, e.g., in hotels, Internet cafes, and amusement parks, where the overall gaming QoE needs to be maximized using already-deployed infrastructures. We also propose two algorithms to solve the gamer-centric problem. We present a prototype system built by off-the-shelf components, and quantify the implication of live migration, which refers to moving a running VM from one physical server to another. We augment our algorithms to accommodate to high migration over-head, resulting in efficient and practical algorithms. Our extensive trace-driven simulations indicate that: (i) our efficient algorithms result in close-to-optimal performance, as small as 0 and 10 percent gaps, (ii) the efficient algorithms scale to large cloud gaming services with twenty thousands of servers and more than 40,000 gamers, and (iii) the efficient algorithms outperform a state-of-the-art algorithm by large, e.g., up to 3.5 times of net profit increase.
You are here: / / Cloud Gaming Experience Towards Placing Virtual Machines Optimizing