Study of web-based management for EAST MDSplus data system

A large distributed data acquisition system based on PXI/PXIe technology has been implemented on EAST tokamak. Now the whole data system has more than 60 data acquisition units and more than 3000 raw channels, and the maximum data throughput is about 5GBytes/s. Most of the acquired data are continuously stored into MDSplus Wan et al. database, moreover the calculated and analyzed data are also saved into MDSplus. In case of long-term discharge mode, the raw data of one shot is more than 1TBytes, and the total data of one campaign is more than 100TBytes. MDSplus is helpful to store tokamak experiment discharge data, and the scientific data are saved into database by shot and tree, and it is easy to know the signal information of one shot or tree. However, it is difficult to get the general metadata information of all the shots and trees. Even if we can read the information by some scripts, but it is not so convenient and the data retrieval speed is slow. So we have planned to design a data statistics and retrieval system for EAST MDSplus data system, which can provide all the metadata information of all the trees such as signal name, sampling rate, data size, data available or not, and so on, basing on these metadata the system can give detailed statistics and reports automatically by some pre-defined model, which can help users to grasp the status and schema of all the data. The system has been developed based on apache/mysql/php framework and the PCIe SSD storage is adopted to promote the database access speed. The design details will be given in the paper.