A new knowledge-based partitioning algorithm in CVE systems

Abstract

Collaborative virtual environment (CVE) system supports large number of users to explore a virtual world and interact with each other through networks, so one of the key issues in the design of scalable CVE systems is the partitioning problem. Existing partitioning algorithms in CVE systems based on multiple-server architecture, in our opinion, hardly consider reducing inter-server messages. In this paper, we propose a new knowledge-based partitioning algorithm which can effectively decrease inter-server messages in CVE systems. In the new knowledge-based partitioning algorithm, using the knowledge matrix built on the communication character of the virtual environment, the cells that have high communication cost are chosen as the centers of regions while the cells that have low communication cost are allocated for the edge of the regions. The experimental results, given in Fig. 6 and Table 2 in the full paper, show preliminarily that the knowledge-based partitioning algorithm does decrease the traffic among the servers in the system and improve the partitioning performance.