Blogger: Lyn Robison
Have you considered a career in modeling? Modeling can be highly rewarding, especially when you are building a data services layer.
When building a data services layer, remember that simply piling on more interfaces to enterprise data sources is not going to provide much value. A jumble of service-based interfaces over enterprise data sources does not make a useful data services layer. The idea behind SOA is to create layers of abstraction, with well-defined schemas and contracts, behind which systems can be loosely coupled.
How do you create layers of abstraction for enterprise data? The basis of an abstraction layer for enterprise data is a logical data model. Logical data modeling is technology-independent, while physical data modeling adapts the logical data models to implementations in physical software systems, such as databases.
With a logical data model as a foundation, you can vary the physical data models as the technology demands and still have well-defined schemas and contracts. If an enterprise has a logical data model that forms the foundation of a data services layer, the physical data sources behind that data services layer to can be refactored.
Refactoring physical data sources can be a very good reason to build a data services layer. A data services layer that decouples software applications from specific physical data sources makes it feasible for you to clean up your enterprise data sources and your enterprise data.
If information is the better half of IT, and I think it is, then logical data modeling is the foundation for an IT department to deliver better information to enterprise businesspeople.