Space-scale structure for information rejection in large-scale distributed virtual environments

Distributed Virtual Reality systems are facing limitations due to the vast amount of streaming communications and world stare update data to handle at host and network level. In order to lessen these burden, good mastering of data flows is required. Host information needs mostly depend on how local observers perceive the virtual world surrounding them. We propose a structure storing a hierarchical description of the world as a basis for host information needs identification. Our information rejection criterion is currently an approximation of the projected area along an axis toward the observer on a perpendicular plane. It can be applied to sources such as visual objects or sound sources. This structure, used to dynamically reference sources based on their position and scale, is called a space-scale structure (SSS); an octree is the actual storing structure. The SSS is traversed in order to retrieve sources, with respect to the rejection function. Optimizations of SSS traversal and SSS update for moving sources, based on temporal coherency, are presented here. A possible distributed architecture using our SSS is also proposed. This architecture makes use of multicast and client/server network topologies. The implementation of the SSS is evaluated locally in the context of large database culling.