RAMM: a game theoretical replica allocation and management mechanism

This paper proposes an agent-based distributed replica allocation and management technique, where each agent maximizes its own benefit, such as, user access time, latency and communication cost. The technique gathers inspiration from market economy and game theoretical mechanism designs. In such mechanisms the agents do not have a global view of the system, which makes the optimization process highly localized. This local optimization may encourage these agents to alter the output of the resource allocation mechanism in their favor and act selfishly. The proposed technique guarantees a global optimal solution even though the system acts in a distributed fashion operated by self-motivated selfish agents. The mechanism is extensively evaluated against some well-known replica placement algorithms such as greedy, branch and bound, game theoretical auctions and genetic algorithms. The experimental results reveal that the mechanism provides excellent solution quality, while maintaining fast execution time.