An algorithm for interest management in High Level Architecture

One of the most important design concerns when implementing large-scale high-level architecture (HLA) simulations is how to perform effective interest management. HLA interest management separates routing from data delivery so as to filter out unnecessary data communication between federated computers. One of the more scalable mechanisms for such interest management, data distributed management (DDM), distributes and filters data based on regional clustering. Essential to DDM, therefore, are the algorithms for dividing routing space and for matching regions. In this article, we describe how the spatial hierarchy of routing space is defined, and introduce a technique based on n-dimensional binary trees to solve the problem of region matching. We then introduce two algorithms: one to compute the coding of hierarchically structured regions, and the other to match regions based on this coding. Finally, we outline our general strategy for data filtering in HLA, and analyze the advantageous and limitations of the algorithm through experiments.