Autonomous massively multiplayer online game operation on unreliable resources

Massively Multiplayer Online Games (MMOGs) are a new type of large-scale distributed applications characterised by seamless virtual worlds in which millions of world-wide players act and interact in real-time. Although for the past decade the number of MMOG players has grown exponentially, to the current tens of millions, this very growth may now hamper the progress of this important branch of the entertainment business. To guarantee Quality of Service (QoS) to a highly variable number of concurrent users, operators statically over-provision a large infrastructure capable of sustaining the game peak load, even though a large portion of the resources is unused most of the time. To address this problem, we propose a Cloud middleware-based system for autonomous operation of MMOGs. Our system provisions resources on-demand from multiple Cloud providers, automatically distributes the MMOG load between these resources, and self-heals when confronted with unforeseen resource failures. This new operational model allows small and medium enterprises to join the competitive MMOG market through near-zero initial infrastructure investment and operate MMOGs at given levels of QoS with small human intervention. We evaluate through simulations based on real-life MMOG traces the impact of resource availability on the QoS offered to the MMOG clients. We find that: (1) our proposed MMOG operation system can mitigate the negative effects of resource failures in under four minutes, (2) MMOG server consolidation in a resource-scarce environment can accentuate the negative effects of resource failures, and (3) the competition for resources can indirectly affect the QoS of the MMOG sessions. © 2013 ACM.