Think about the various mobile phone apps that attach to your calendar. They don’t own your calendar but they know of it and can schedule events on it, set reminders, etc.
The data landscape of an enterprise is more complex than your calendar, but not impossibly so. Experience shows that it is possible to establish a shared model of the essential concepts that underpin all of the applications of a large firm.
The data landscape becomes an onion-like structure with raw or semi-processed data (think of a data lake or a triple store) and an ontology (think high-level data model) to interpret it and a guard (a layer that authorizes use and prevents bad things from happening to the data).
Once established, becomes the common denominator for the apps that perform the work of the firm.
The Data-Centric revolution will not be application-driven. There will be no killer app. That’s the point.
|NOW: Application-Centric||FUTURE: Data-Centric|
|Exorbitant, often prohibitive, cost of change.||Reasonable cost of change.|
|Data is tied up in applications because applications own data.||Data is an open resource that outlives any given application.|
|Every new project comes with a big data conversion project.||Every new project taps into existing data stores.|
|Data exists in wide variety of heterogeneous formats, structures, meaning, and terminology.||Data is globally integrated sharing a common meaning, being exported from a common source into any needed format.|
|Data integration consumes 35%-65% of IT budget.||Data integration will be nearly free.|
|Hard or impossible to integrate external data with internal data.||Internal and external data readily integrated.|