HLA federate migration

The high level architecture (HLA) is a standardized framework for distributed simulation that promotes reuse and interoperability of simulation components (federates). Federates are processes which communicate with each other in the simulation via the run time infrastructure (RTI). When running a large scale simulation over many nodes/workstations, some may get more workload than others. To run the simulation as efficiently as possible, the workload should be uniformly distributed over the nodes. Current RTI implementations are very static, and do not allow any load balancing. Load balancing of a HLA federation can be achieved by scheduling new federates on the node with least load and migrating executing federates from a highly loaded node to a lightly loaded node. Process migration has been a topic of research for many years, but not within the context of HLA. This paper focuses on process migration within the HLA framework.