The IPOPP Data System (DS) Control System (CS) controls product generation. The CS acts as the interface between the Data Storage Manager (DSM) and the Science Processing Algorithms (SPAs). The DSM maintains product storage. The CS retrieves SPA input products from the DSM, and creates the environment for execution of the SPAs. Once the CS has executed an SPA, it registers the resulting output products with the DSM.

The CS is data-driven and contains any number of stations, each of which controls an SPA. Each CS station has a control program and a subdirectory. When started, each control program reads the configuration file to determine the resources required by the SPA and requests these resources from the DSM. Resources may consist of ancillary data or previously generated ancestral products. When the requested resources are available, the CS station creates the required environment NISFES and executes the SPA. The CS station then informs the DSM of the new product, removes the environment, and again queries the DSM for requisite resources to make a new product. Multiple CS stations may run the same algorithm type in parallel to speed execution. These stations may reside on different machines in order to balance the processing load.

The CS architecture is decentralized and based on the notion of a station. Each station runs an SPA when the required resources are available. There may be any number of stations in the CS, as depicted in the "CS System Description" diagram. The CS design incorporates the following features:

• scalability, to accommodate indefinitely increasing processing loads;
• language-agnostic operations (i.e., SPA software may be written in any language);
• automatic execution of programs in proper order; NISGS Front End (NISFES)
• product generation restart;
• easy addition of new SPAS;
• rapid fault isolation;
• platform_independence