摘要：Inconsistency is a challenge with many delaysensitive Massively Multiplayer Online Games (MMOGs). In this paper, a framework for inconsistency analysis and control is proposed. A typical situation, called hot-spot region, is considered for resolving this problem, since it is one of major causes of the inconsistency. Stochastic models are presented to abstract the operations of a MMOG in the context of our goal, particularly for examining the impact of various system characteristics and player behaviors on the inconsistency. A definition of game inconsistency together with a method for computing inconsistency rate is also provided. By using our models, we carry out inconsistency analysis quantitatively and obtain a deep insight into its behavior. Based on our analysis, we develop a mechanism for inconsistency control in the context of hot-spot region, so as to illustrate our approach in this area.
关键词：massive multiplayer online games, federated MMOG p2p architecture, queuing theory and Markov processes, performance modeling.
