Adaptive load partitioning algorithm for massively multiplayer online games

Abstract

Distributed Virtual Environment Systems are widely used in massively multiplayer online games. With an efficient distributed architecture and load balancing algorithm, they can support tens of thousands of players interacting with each other. An existing prevalent mechanism is to divide the virtual world into several regions and microcells and use the graph-based partitioning algorithm. But they generally limit the assignment beginning with border nodes and can't adapt to the situation when players are crowded around the center of the entire map. Furthermore, many algorithms simply focus on reducing the intra-server communication cost in a connected graph. In this paper, we propose an adaptive load balancing algorithm to solve the problem of the center-crowdedness in the DVE systems, while reducing the intra-server communication and client migration cost simultaneously. © 2013 Springer Science+Business Media.