Abstract
Using Data Distribution Management (DDM), federates can send/receive interest data properly instead of broadcasting data over entire network. For large-scale HLA simulations with many simulation entities or federates, the use of DDM is essential to support such simulations. However, the scalability issue of DDM is rarely raised in the literature. Thus we present a scalable DDM approach, DHT-DDM, using Content-Addressable Networks (CANs) [1] based on Distributed Hash Table (DHT). Region declaration, region matching, and network connection mechanisms for DDM are designed considering scalability. The declaration of regions is done deterministically without depending on global information, the load of region matching is distributed and shared among federates, and the connections of federates is built upon application-level multicast.