A novel compatible runtime infrastructure under complex environment

Originating in 1980s, distributed simulation has turned out to be a powerful tool in scientific research, military simulation and training, complex product design and more other fields. However, new problems like limited computational capacity and heavy communication overhead exist as simulation modeling and computation develop. This paper concentrates on these problems, reviews related work, and proposes two novel ways to improve traditional RTI (runtime infrastructure) under new application scenarios (High performance cluster and reflective memory network). The proposed methods of combining software improvement and hardware enhancement are demonstrated to be effective and efficient.