Data interoperability via surrogate API libraries

This paper presents how a client-server system can be used to create an efficient and interoperable data delivery system for accessing local and remote data through various existing data analysis tools. The system is based on a a series of servers that provide data over the network through a data access protocol (DAP). Each server will read the data from its original format and deliver it to the client according to the DAP specifications. New client software can be developed around this DAP and existing software packages can become a client by a re-link. During the re-link the standard application programmer interface (API) library is substituted with a surrogate API library which has been developed with the DAP. Therefore, the translation, selection, extraction, and transport of the data is hidden from the client software. Basically, each client software sees the data available from all the servers as local files in its native format where the file names are changed to a fully qualified Uniform Resource Locator (URL). The fully qualified URL adds power and flexibility to the system by not only simplifying where to access the data, but also by providing a general mechanism for passing arguments to the server side functions which extract, process, and convert the data on the server before delivering it to the client.