Performance analysis of parallel data distribution management in large-scale battlefield simulation

There are so many operational entities in large-scale battlefield simulation. Entities may interact with each other in a real-time and frequent manner, which may restrain the performance of the simulation system. From researches in recent decades, it can be concluded that the Data Distribution Management (DDM) is an effective solution for this issue. Although the main focus for realizing DDM is Pub/Sub paradigm, recent literatures has implied that the serial realization of various DDM algorithms has encountered the performance bottleneck. The development of computer microprocessors is making progress from multi-core to many-core, so developing the parallel processing capabilities for DDM in multi-processors platform is the most effective approach to promote runtime performance. Based on the analysis and comparison of region-based, fixed grid-based, dynamic grid-based and hierarchical grid-based algorithm for region matching, the parallel algorithms for these methods are realized according to the matching task decomposing. The experiments results manifest that parallel algorithms can gain about 40%~60% performance improvements and support the research of parallel simulation engine of wargame.