Using projection aggregations to support scalability in distributed simulation

Abstract

Distributed interactive simulation systems are growing to include well over 100,000 dynamic entities for applications such as multiplayer video games, military and industrial training, and collaborative engineering. In these applications, each host receives updates (such as position and orientation) from remote entities, models and renders the scene, and pegorms other tasks such as collision detection. The number of entities places a heavy burden on both the networking resources and computational resources available to the application. To address these limitations, some systems have aggregated information about groups of simulation entities according to their organizational structure or their location within the virtual world. Howevel; traditional aggregation techniques are inadequate because remote hosts need to access entities based on both their organization and their virtual world position. This paper describes projection aggregations, a technique for grouping entities by both their organization and location. Remote hosts use projections to control which entities are represented locally and at what level-of-detail. We describe how projection aggregations are implemented in a networked environment and demonstrate how they reduce network bandwidth and computational requirements, Finally, we argue that projection aggregations represent a general-purpose framework for representing all simulation entities, thereby supporting evolution of entity models.

Survey of state melding in virtual worlds

“…In the Paradise project [Singhal and Cheriton 1996], projection aggregation is applied in which each projection aggregation includes objects from a single organization (e.g., a tank that consists of wheels, its body, treads, turret, etc.) located within a single region.…”

Crowded collaborative virtual environments

“…The Paradise project has been exploring the use of graphical abstraction techniques to produce aggregate views of groups of objects for use in distributed simulation [6]. Their approach generates statistics about graphical objects (e.g., the number present, their mean location and spread) and these aggregations are then used to generate graphical representations.…”

Component framework infrastructure for virtual environments

“…Some of the most well known systems supporting LSVE are listed: AVIARY [25], BrickNet [23], DIVE [14], MASSIVE [12], NPSNET [17], PARADISE [24], RING [11], SPLINE [1] and VLNET [6]. These systems focus on particular applications, leveraging whenever possible the particularities of each problem domain and consequently constraining the design space and consequently reducing the overall complexity of the implementation.…”