Today's simulators involve many varied computational systems. Interoperating these devices is a key strategy for extending the life of current simulators. The Distributed Interactive Simulation (DIS) protocols solve the maze of interoperability.

These statements are well-known to the simulation industry, but the stark reality of the situation is that varied systems produce varied implementations of DIS. The full gamut of DIS involves the varied processors, varied operating systems, varied programming languages and structures, varied interface hardware, varied coordinate system implementations, and varied data base formats. Realizing these six areas simultaneously is a particularly demanding chore. This paper attempts to show how one implementor went about producing code for DIS, seeking to provide reusable code in the process. The lessons learned from this venture are discussed.

Using a PC-based radar simulation system as the baseline, the paper discusses the research and development of DIS in this varied environment. Although a radar appears to be a "receive-only" entity on a DIS network, in order to locally test such a system, test vectors (or PDUs in the DIS parlance) must be generated. Thus, the baseline requires some way to construct test vectors, such as through semi-automated forces (SAFOR) or Computer Generated Forces (CGF) generators. A limited CGF for the required purposes is described in the paper.

Other steps in the implementation include actually producing, transmitting, receiving, and displaying the CGF state vectors. Production involves coordinate conversion schemes, PDU receive and transmit functions become as dissimilar as their associated processors, and display techniques require limitations in the scope of what can be displayed. So, the paper surveys network I/O techniques and selects the correct one for the radar simulation. The last stage (displaying) requires a filter of the vectors since oil processors (and especially the PC in question) have limits in terms of computing power.

Intermediate steps in the full implementation of a DIS system involve determination of correct protocols sent from the CGF and the use of terrain and feature data bases. Both of these areas are also discussed including the fields of network analyzers, DMA maps, and Project 2851 SIF.

The paper points out that, although realizing a DIS interoperation can be straightforwardly done, care must be taken to understand that there is more to the "varied" problem than just the obvious processor incompatibilities.

Menu

John O'Reilly