What is OODT?

If you've been scratching your head with questions like "Middleware for metadata? What the heck does that mean? Can you just give me an example?" you're in luck. Read on for a quick and dirty example of what OODT is and what it can do.

An Example

OK, suppose Fred Farnham at Farnsworth Technical College records the speed of a prototype particle in a vacuum in furlongs per day. He's got an entire table of results tabulated from his testbed and stored in a flat file on an EBCDIC system. And suppose Peggy Portoluca at the Proboscis Institute records the speed of the prototype particle in a vacuum in miles per hour. She's got a disk full of data stored in a dynamic database.

And you want do some research on the speeds of prototype particles, preferably in meters per second. No databases. No EBCDIC. Just a nice standard format of all the results. OODT lets you do just that.

OODT's profile service describes the locations of available information. Using metadata, OODT can manage the correlation between furlongs per day and miles per hour. It profiles electronic resources' composition, including as much or as little information as required. Synonyms between data elements, ranges of values captured, units, data types, and more are part of the metadata schema that the profile service uses. The profile service describes any electronic resource, including instances of the product service.

The product service is another OODT component. Using the identical query structure as the profile service, it links to institute-specific databases, flat files, or other proprietary storage, locates matching products, and converts them i nto Internet-standard formats.

Other components include the query service which controls the propagation of queries throughout a web of profile and product servers, and the archive service which archives newly created and/or ingested products along with metadata into a catalog and repository.

OODT is a software framework to support metadata (profile) and data (product) search and retrieval. It's written in Java. It has client APIs in C, Java, and Interactive Data Language, as well as supports HTTP access, enabling easy integration into web-based applications. Combining a simple installation process (and a self-installer on Windows platforms) with remote management capabilities makes it easy to deploy far and wide.