Multi-objective zone mapping in large-scale distributed virtual environments

In large-scale distributed virtual environments (DVEs), the NP-hard zone mapping problem concerns how to assign distinct zones of the virtual world to a number of distributed servers to improve overall interactivity. Previously, this problem has been formulated as a single objective optimization problem, in which the objective is to minimize the total number of clients that are without QoS. This approach may cause considerable network traffic and processing overhead, as a large number of zones may need to be migrated across servers. In this paper, we introduce a multi-objective approach to the zone mapping problem, in which both the total number of clients without QoS and the migration overhead are considered. To this end, we have proposed several new algorithms based on meta-heuristics such as local search and multi-objective evolutionary optimization techniques. Extensive simulation studies have been conducted with realistic network latency data obtained from real measurements using millions of pairs of geographically distributed IP addresses, and different workload distribution models. Simulation results demonstrate the effectiveness of the newly proposed algorithms.