The central goal of data architecture is to enable the capability to learn the truth. As a result of data architecture all data that can be in scope is addressed, even if some of the data is unavailable.
Data is the accumulated set of values during the course of some activity.
Information is the use of data in a context.
Knowledge is the ability to understand the implications of the data and information.
Wisdom is the independent ability to decide given the entire universe of available and unavailable data, information and knowledge.
If we navigate data using an architecture, the data can be used in meaningful ways, depending on the objectives. The ability to gain knowledge and wisdom can be enhanced significantly with an architectural approach.
For example, one of the objectives of data is to address a business purpose. Let us say we want to find out our most valuable customers. The answer to this may be as simple as looking up the customers and our sales and profit from them and ranking them as needed based on our definition of valuable. On the other hand, it may be difficult to impossible to find this information without a substantial amount of effort.
It is likely that a data architecture would streamline the approach to understanding data to mitigate the difficulties in answering such questions.
How can it be complex?
What appears to be a simple question in the business perspective is suddenly a complex project from a data perspective. The complexity arises from the lack of confidence in the array of data sources, data structures combined with dissimilar terminology to address common data elements. The more diversity one encounters in these aspects, the more complex it gets.
What is the solution?
Align the data elements and data in ways such that the diversity encountered can be controlled to transform it into a consistent view. This may be a inordinate amount of work, but that is the approach one might take to meet the business purpose.
But WHAT is the solution?
Ideally, we should have prevented this in the first place. If answering such questions was effortless, business could proceed at the speed of thought.
So, what is our solution?
At this time, the leading thought points to the creation of a data architecture. A data architecture enables the classification of data elements into various hierarchies, minimally one hierarchy that provides a taxonomy for the purposeful navigation of the data. Such a taxonomy allows data to be managed in accordance with its purpose and classification and not merely due to its presence or absence.
The Data Reference Model in the Federal Enterprise Architecture provides guidance on approaches to data architecture that can serve multiple purposes. Refer to XXXXXX for more details.
The role of data architecture in Virtually Managed Software(R) is fundamental. Using this approach, all data is modeled in accordance with the guidelines of a fundamental organization that is independent of its interpretation. Conceptual data models are still interpretations. Logical and physical data models are too specific to be independent of their interpretation, as they represent composite structures, and hence subject to change. It is easier to understand why this is, because logical models are consequent to requirements, and physical models are consequent to their logical models. Thus, their interpretation is tied to the requirements that originated them.
Recognizing the need for Data architecture is the first step before arriving at a virtually managed data model. It is not only independent of its physical and logical design and implementation, but even independent of any particular interpretation. Virtually Managed data models are architecture based, yet support multiple interpretations.
It is therefore clear that a data architecture is fundamental to the ability to enable any interpretation of data, and architecture goes beyond modeling of the data. If an architecture centric approach is used to address data, then data can be more than useful – it can be reused in multiple situations with effortless ease. The Virtually Managed Data Model is the key to enabling such a data architecture.