دسترسی به لایه داده های استاندارد برای دستگاه ترکیب در برنامه های تحقیق وتوسعه

Each EURATOM Association stores data using proprietary schemes, usually developed by the research unit or using third party software. The temporary exchange of researchers between laboratories is a common practice nowadays. When the researchers returns to the home laboratory, there is usually the need to follow the work. The amount of available data is becoming enormous and the main data index is changing from shot number to time and events, where the pulse number is just one among the most relevant events against data is catalogued. These difficulties can be overcome by using a common software layer between end-users and laboratories. The components needed to create this software abstraction layer, between users and laboratories data, have already been developed using a universal and well known remote procedure call standard (RPC) based on the eXtensible Markup Language (XML): XML-RPC. The library allows data retrieval using the same methods for all associations. Users are authenticated through the PAPI system (http://papi.rediris.es), allowing each organization to use its own authentication schema. Presently there are libraries and server implementations in Java and C++. These libraries have been included and tested in some of the most common data analysis programs such as MatLab and IDL. The system is already being used in ISTTOK/PT and CASTOR/CZ.

With the internationalisation of science it is very common that scientists need to access data from different laboratories. Each laboratory has its own way of retrieving data and users need to adapt themselves. Instead of concentrating in the data analysis, scientistshave to spend time and effort learning howthe different data access schemes work, change their analysis codes for each experiment and manage an updated version for the different programs and libraries which are required to fetch data in the different scenarios. To solve this problem the best approach is to hide all the complexity from end-users, so that users only have to learn once how to access data and to maintain a single updated version of libraries and software. This does not mean that every association must store and retrieve data in the same way, not limiting therefore the test of new technologies and new data stores solutions and algorithms. The key point is that scientists must be as isolated as possible from different IT (information technology) tests and solutions and should only have to worry about data analysis and results. This paper presents a solution that implements the ideas described above.

The Shared Data Access System allows scientists to focus only on data analysis without having to worry about data access procedures. Moreover, the user codes do not need to be changed to fit the host laboratory needs. There is native support for integrating the SDAS libraries into the most commonly used data analysis programs like MatLab, Mathematica and IDL. The systemwas designed to be as generic as possible and can be adapted to any data storage mechanism. In the near future bi-directional support for data will be added, allowing users to submit processed data. SDAS provides a valid alternative for data indexing. Using events and time-stamps the system is prepared for the challenges of the near future, maintaining backward compatibility with the solutions of the past. Currently, this data access solution is successfully being used at ISTTOK/PT and CASTOR/CZ [8].