Database Technologies On The DIS/HLA Simulation Run-Time Environment

It's very important for simulation systems to manage data effectively during the process of simulation modeling, simulation execution and after-action analysis. Especially for the simulation execution, the system often needs to access current data in a real-time manner. In order to provide the users with statistical information as soon as possible, the system also needs to manage historical data effciently. The temporal database is a database system that can manage time-varying information of the world effectively, and it can also meet the users' needs to access historical and current data in a consistent way. As a result of very large amount of data being managed, the storage and query technologies are difficult and important research issues for temporal databases. So the real-time management technology for temporal data on the DIS/HLA run-time environment is deeply studied in this paper.Data collection system is a typical component which is closely related to data management in the DIS/HLA simulation system. The storage architecture of collected data, the strategy of data loading and the usage of collected data are first studied in this paper. Then a log-based temporal data model is provided to deal with time varying information of the simulation run-time environment. Only a time attribute is used as the temporal attribute in this model. On the base of the relational algebra of the model, the simulation system can query the historical data during the simulation execution effectively and analysis the simulation data after simulation quickly. The snapshot querying and schema transformation algorithm defined for the model can also be used by simulation systems to implement the simulation reviewing and data format exchanging of different simulation architectures. Experiments show that the efficiency of data storing and querying in this model is better than that of file system and SQLServer system.According to the data model, two new access methods (indexes) used for simulation run-time environment are designed in this paper. They are the Quasi-Equal Interval Sparse Index and the Variable Hash Index. The former is an ordered main-memory index used for sequential file, and the space of the index structure was constant. Analysis and experiment rusults show that the index can greatly reduce the times of I/Os without affecting the data collection process. The latter is a dynamic Hash index, and the temporal attribute is used as a Hash variable. The index is well suited for time attribute because of the idea of adaptively changing of Hash function parameters, and those parameters are organized as an AP-tree. It can...