What is a Data Fabric?

Data fabric has been named as one of Gartner’s strategic technology trends for the coming year, but is its inclusion in the analyst’s predictions for 2022 is a little premature? Here, Euler CEO, Rob Jones, covers what a data fabric is, how it can be used, and why he feels we’re still some years away from Gartner’s vision of it becoming a genuine ‘trend’.  

What do you mean by data fabric? 

While data fabric itself isn’t new, the way it’s evolved as a concept has changed dramatically over the last two years. It was previously considered as an array of tools that were used to handle all kinds of data processes – from discovery, cataloguing and preparation, to automation, transformation, stewardship, master data management and more. Today a data fabric is even more advanced, using AI and machine learning to create an intelligent always-on, always evolving service that identifies issues before you can, and fixes them automatically using the intelligence it’s built up. 

So how does it exist on the market today? 

The challenge with the concept of a data fabric is that everybody’s data, architecture, and mix of technologies is different – there’s no ‘one size fits all’. There are not currently any off-the-shelf tools to provide Gartner’s vision of a data fabric. Even Gartner themselves admit that the current tools on the market provide around 50 to 60% of the foundation, but they’d still require you to build the AI and machine learning elements yourself. And even then, the likelihood is that organisations would need external expertise to deliver this. It requires the skills to select and blend the right technologies and to build and customise the AI applications. It would require you to select the machine learning language or AI platform you want to use and then write all the coding to collate and integrate the different elements. 

How far away do you think we are from a single, genuine ‘off the shelf’ data fabric product?   

I’d estimate that we’re still three to four years away from that. It’s likely that the first all-in-one solution will eventually come from a big cloud vendor. And even then, it’s likely to be very specific to their cloud and would therefore still require configurations to make it fit for purpose. The level of expertise needed to deploy it would be incredibly high, so it would still most likely be a job for a data specialist.  

How would it be used in organisations?  

As it stands, big organisations tend to employ tens of people to manually correct, manage and maintain data integrations, distribution, and governance. A data fabric with AI and machine learning could automate these processes and ease the burden. It would still require involvement from people within the organisation to help train the AI, but would mean you could significantly reduce any existing headcount. This would be especially beneficial in today’s challenging recruitment market as we’re facing a huge shortage in data management and integration skills. Any data skills that organisations manage to free up as a result could be redeployed to focus on other urgent tasks.  

An effective data fabric also has the potential to spot and correct minor semantic issues with your data. The ones that you wouldn’t normally be able to see until they become major problems that required significant manual intervention.  

What’s your advice to organisations thinking about investing in a data fabric in the next few years? 

The rewards that data fabric can bring in terms of efficiency, accuracy, and capability, particularly if you’re a large organisation, are huge. However, there’s still not a lot of expertise around the new definition of a data fabric, even in the supplier sector, because it’s so new. You’re going to have to expect that if you’re looking into this, you’re a trailblazer and you’re going to make mistakes. Even though you might be a leader in the adoption of data fabric, you might not quite meet the business case because it’s still such an unknown area.  

While data fabric itself isn’t at an advanced enough stage for mass adoption through 2022, when organisations are acquiring new data management technologies next year and beyond, they need to have an eye on how they’re going to be compatible with a future data fabric.  

To find out how we can help with data management and integration, or for more information, simply get in touch.


Related Articles

Data Distrust – The impact it has on businesses and how to eliminate it

15 Sep 2021

Poor quality data can affect many different areas of your business. Efficiency can be stifled, reporting inaccurate,…

Why a data migration framework will help your project to succeed

10 Dec 2021

The most recent figures suggest that nearly 40% of data migration projects end up being over deadline,…

Why data integration tools should be on your radar

29 Nov 2021

How clear and consistent is the data your organisation relies on for decision making and planning? Is it…