Techniques for handling scale and distribution in virtual worlds

Lack of bandwidth and network latency are known to be major impediments to achieving realism in distributed virtual world (vw) applications with a large number of, potentially geographically dispersed, entities. This paper describes a combination of techniques that we are using to overcome these twin problems. The techniques described are intended to reduce both the volume and frequency of communication between the entities that make up the virtual world and include the use of anonymous event-based communication with notify constraints, scoping of event propagation with zones, and use of predictive approaches to replica management. Each of these techniques is described in turn.