Building a Data Mesh Architecture in Azure – Part 16 – Dead or Alive?
Is the Data Mesh dead or have we been using it without realising?
The Gartner Hype Cycle for Data Management 2022 curve (shown below) is now a well-known picture in certain circles of the data community. It seems the trolls love to point out the relatively low position that Data Mesh has on this curve and the red cross key advising: Obsolete before plateau.
This view then gets accompanied by posts claiming the “Data Mesh is dead; it just doesn’t know it yet”.
Hence, I felt like adding my opinion to the debate, as follows.
In opening… You fools! 🙂
The Data Mesh is far from dead.
In fact, you have already been using such an architecture with great success for many years.
Hint, its behind you!
Yes, lets ignore the hype and realise the fruits of a de-centralised platform.
What platform am I talking about?
What ‘data mesh’ like architecture have we been using already?
The best example of a de-centralised platform.
Literally organised into domains.
Web pages as data products.
All available via a ‘mesh supervision plane’ that we know as our favourite search engine. Google in my case.
Query this de-centralised mesh of knowledge and information that is the internet via this frontend tool (Google), with support natural language, and question/answer intent.
Data search capabilities available in many forms and distilled as information available by human interface devices such as Alexa!
Maybe we could even say the common web browser is the ultimate mesh frontend tool.
Data/information/knowledge all available to non-technical citizen users. Like my Grandad!
Federated governance in the form of TCP/IP. DNS. HTTP. SSL. FTP. REST. etc etc.
The Data Mesh isn’t dead.
We have already created it for global public datasets and information sharing.
Let’s create it for our private, closed datasets that fuel our businesses.
Google and Alexa for your enterprise datasets. Available to everyone in the company to self-serve.
Apologies, this post became a little more of a rant than I was intending. Hopefully, you can excuse this as a reflection of the passion I have for what I do day-to-day and ultimately share my point of view. Or at least can agree the data mesh isn’t dead.
It’s hard to deliver. But that does not make it dead.
Many thanks for reading.