Supporting large-scale networked virtual environments

In this paper, we introduce a framework that supports a large number of participants in a collaborative virtual environment (CVE). We consider object classification based on avatars' physical characteristics to define the overlay network used to disseminate update messages between participants. This concept systematically puts objects into different logical clusters where each cluster can be rated in terms of a persistency class. We integrate a greedy heuristic algorithm to form a better network structure and then apply the dominating set principle to build a routing table for intra-zone communication. In this paper, we have used OPNET to verify the applicability of these approaches to large-scale CVEs.