Enterprise Semantics: Speed-Reading Your Enterprise Data Architecture, Part I

by Ken Orr

Every large enterprise has what I call "one huge accidentally distributed database"; i.e., thousands of individual database tables, files, spreadsheets, and document stores with the same data elements stored redundantly all over the place and multiple copies of that same data stored over and over again. All of this complexity takes time to get your brain around if you're a new enterprise data architect. This is made even more difficult because the names for the same data elements are not always (or even often) the same. Finally, throw in the fact that the data often is incommensurate, in that it represents different time groupings, and you have a very tough terrain to map. It is absolutely necessary from an Enterprise Architecture standpoint, however, that the data architecture folks understand it.