Efficient Zone Mapping Algorithms for Distributed Virtual Environments

摘要

This paper deals with the zone mapping problem in large-scale distributed virtual environments (DVEs), e.g., massively multi-player online games, distributed military simulations, etc. To support such large-scale DVEs with real-time interactions among thousands of concurrent, geographically separated clients, a distributed server infrastructure is generally needed, and the virtual world can be partitioned into several distinct zones to distribute the load among the servers. The NP-hard zone mapping problem concerns how to assign the zones of the virtual world to a number of distributed servers to improve interactivity. In this paper, we propose new zone mapping algorithms based on a Linear Programming relaxation of the original problem and meta-heuristics such as local search and evolutionary optimization techniques. We conducted extensive experiments with realistic Internet latency models obtained from real measurements using millions of pairs of geographically distributed IP addresses. The results have shown that our newly proposed algorithms significantly improved the performance of large-scale DVEs in terms of overall interactivity, when compared with existing mechanisms.