Research On Key Technologies Of Grid-Based High Level Architecture For Simulation

LAN based tranditional HLA can't support the mass and complex distributed simulation, lack of flexibility, scalability and resource reusability. Simulation Grid, the combination of Grid and Modling&Simulation, can achieve the share and reuse of resource in simulation, cooperation between different simulation, dynamic scheduling, fault tolerance and load balancing. We can solve the issues of HLA limitations by applying grid technology.Grid based Distributed Simulation Platform (GDSP) is the base infrastructure of Service-Oriented extensible Modeling/Simulation supporting environment architecture. The Grid-Based HLA comprise two important Grid services in GDSP, Grid-based RTI(GRTI) and Manager Center. RTI is the software implementing HLA Interface Specification.The two services can promote reusability and Scalability of GDSP, better reuse existing simulation resources, and obtain more effective management of resources.An Agent-Based Call&Callback Strategy in GRTI and SFN-based RTI Scheduleing algorithm in Manager Center is adopted. GRTI service implements the HLA interface specification. GRTI encapsulation a traditionanl RTI -CRTI, so can support the GRTI version and CRTI version application to run in a same federation. Manager Center manager the resources in GDSP, RTIs, federations, federates and simulation models in different version.Those resource can register to and deregister from it. SFN-based RTI Scheduleing algorithm is used to daynamicly schedule RTI for different federation.A Grid based simulation model specification in GDSP was introduced. It defines the interface and executing flow of simulation model in GDSP. Simulation models in GDSP are Grid Servicse,can be written in different programming language, deployed in different OS and reused in different simulation application to reduce the workload and complexity of simulation project.Through the system test, Grid-Based HLA can support the Grid-based simulation, conform the simulation resource in GDSP, schedule RTI for simulation, achieve load balancing, and improve the whole performance of GDSP.