A Distributed Data Architecture for 2001 Mars Odyssey Data Distribution

The Planetary Data System (PDS) has had a long tradition of using metadata. Metadata, which is literally data about data, describes the shape and meaning of data. Metadata enables both human researchers and automated systems to determine if a certain piece of data is the right piece, what operations may be done on it, whether it is of value for the current course of research, how it relates to other data, and so forth. What the PDS lacked was the middleware to handle that metadata [8]. Middleware, which is literally software components that glue backend and frontend systems together, was exactly what the PDS needed. The Object Oriented Data Technology (OODT) task at NASA’s Jet Propulsion Laboratory had been developing a software framework for handling data and metadata in a uniform way. This framework acts as a middleware system, enabling backend data systems to plug in and share their data with frontend systems providing access. The middleware encapsulates differing representations, formats, locations, and meanings of data, making data interoperable and relieving researchers of requiring foreknowledge of localized data system organization and representation. Using the OODT Framework, PDS developed a nextgeneration data distribution system that was online and web accessible, transcending the traditional CD-ROM distribution model [7].