Application Of P2P In EAST Data System

The Experimental Advanced Superconducting Tokamak has constructed at the Institute of Plasma Physics, Chinese Academy of Science. A computer data system– the EAST Data System– has being designed to support the physics research activities by providing diagnostic data acquisition and data analysis. The system is a distributed computer system with a high degree of modularity and capability for autonomous operation. The previous design for the EAST data system (named HT-7U data system) was based on C/S mode. But with the improvement of EAST experiment, the C/S-based EAST data system has been unable to meet the requirements of experiments. So we need to design new data system to improve system performance.A P2P-based EAST Data System will provide data acquisition and analysis support for the EAST superconducting tokamak. In the P2P environment, both online and offline systems are based on Napster-like P2P architecture. In the course of the experiment, all collected data are stored in the data acquisition subsystems locally, and then transmit to the data servers. On the online systerm, the index server is uesed to manage index table for all raw data. PC clients can access the raw data directly using the P2P architecture. This can significantly reduce the completion time of raw data display and real-time processing. On the offline systerm, the index server only manage a few index information, and PC Client act both as a client and as a server . These improve the efficiency of the system.We designed the EASTNap and EASTScope in Linux and Windows environments respectively, and use poll function to manage connections. The new designed data system are implemented and used on the existing HT-7 Tokamak for testing. An analysis technique and a simulation-based method are used for performance evaluation. Comparing the P2P-based system with the C/S-based system, both simulation and analysis studies indicate that the new designed P2P-based system can significantly reduce the completion time of raw data display and real-time processing of online system, and raise the workload capacity and reduce the average delay of the offline system.