What is a
To grasp the value, let's explain the foundation.
Born with the "stitching"
The majority of companies have a common challenge with their data. As companies invest in more innovation, we are often left with an operational debt in the form of data silos. For leaders to get a full view of the business, it is often made impossible due to these data silos and the "Data Fabric" aims to solve this challenge. The data industry has spoken and it is clear that there are some common pillars of Data Management that are needed so that the problem can be solved. These "common pillars" are the "Data Fabric" - only, this time, they are not different products, but rather one platform that encapsulates the core pillars in a consistent and robust way.
Let's take data to 10 6.5
In order for the "Data Fabric" to be ubiquitously accessed it needs to have your data in a state that is mature, but not too mature. Quite simply, the data needs to be in the highest fidelity and most "useable" state it can possibly be in. It is tempting to think that this data exists within source systems where the data is maintained, but this is simply not true. The goal of the Data Fabric is to offer access to your data where you should need to do as minimal work as possible to get it operational in adding value.
You need to set a foundation of high fidelity
As an industry, we need to get away from the thought of solving everything in our data landscape with JUST a relational or derivative of a relational database. Although critical in the technology stack, if our goal is to offer the highest fidelity, highest flexiblity and highest chance for reuse of data then you need to base your data off the highest fidelity database we have available to us today - the graph. Even with this in mind, we need to make sure that we are not "bending" databases to do something that they cannot, should not and are not great at. This is why a Data Fabric must be based off a core set of different database technologies as to support the flexiblity needed to fuel a business with data that is ready to use.