We were excited to see Philip Howard's piece on the importance of a common data model. It was heartening to read our convictions about IT architecture espoused by someone with no overt commercial agenda. Though he mentions our favorite product at the end of it, we did not solicit or contribute to the piece. So we interpret his research as reporting on an industry trend where we are at the vanguard.
That bit of bragging aside, being ahead of the trend means we'd like to see a common data model become more common. To many IT folks "Enterprise SOA" still only equates to normalized interfaces and messaging.
For many years we've seen diagrams with squares, cylinders and computers on either side of and connected to a fat double-headed arrow. The squares, cylinders and computers represent processes, databases and applications. The fat double-headed arrow is labeled as some kind of bus or backplane. Ten-plus years ago it might've been labeled "Object Bus" and the underlying technology might have been CORBA. Five-plus years ago it might've said JMS or other messaging protocol. More recently the fat double-headed arrow just says ESB.
Common messaging protocols (e.g. JMS) and interface definitions (e.g. WSDL) are critical to Enterprise SOA. But messaging and interfaces focus on normalizing business processes and do little to normalize the other half of computing: the data. That is where the common data model comes in.
SOA What? A common data model (sometimes called a common information model) is the data complement of an enterprise service bus. It is a critical component of the fat double-headed arrow.