Using compensating reconfiguration to maintain military distributed simulations
Donald Welch, James M. Purtilo · 1997
In Department of Defense (DOD) distributed training simulations, events occur that can cause unrealistic behav,ior.These events can be system events, such as the loss of a host or a network link to that host.They can also be events that happen only in the simulated world, such as an activity that migrates, bringing increased activity to a different federate.As a result of these events, the distributed simulation may have to restructure itself to maintain realistic behavior.The restructuring must take pla.ce in accordance with a set of rules mandated by both the domain and specific application.Restructuring decisions must rely on information from both the simulated world and the system configuration.Reconfigurations must be made quickly to minimize unrealistic behavior.Automatic reconfiguration is not only faster than manual, but automatic restructuring brings the added benefit of fewer support staff.We call the automatic restructuring of a distributed application with respect to at set of rules Compensating Reconfiguration and we have developed a software engineering environment that could be used to support its inclusion in DOD distributed simulations. INTR.ODUCTIONThe Department of Defense (DOD) distributed simulation domain encompasses a wide variety of uses, architectures, and techniques.DOD uses distributed simulation for test and evaluation, analysis, and training.Each of these categories place different requirements on the dist#ributed archit#ecture.Currently DOD has simulations that use a totally distributed approach, as discussed in DIS (1994), and Weatherly (199G) but has mandated that all simulations use a middleware approach as defined by the high level architecture (HLA) discussed in DMSO (1997).HLA is designed to support a family of simulations that in-