Data Architecture – An effective approach

 

 

This will be a short article, as it is Sunday and my family deserve my attention.

I have seen so many absurd ideas about true data architecture – most of which consider physical approaches as if the architect is a dead role – it is not – and the consistent failure in projects I have been witnessing in most large companies confirm that. People are so concerned about production issues that they simply neglect planning and start believing in developers and product specialists.

Agile practices are blinding people to the fact that simple planning and a bit of design can do wonders for data projects. Against immediatism – the greatest responsible for multibillion losses. While companies impose a distance between IT and business there will be (large) losses – we are not operational expenses.

Besides, there are the information silos – these hidden reigns of information that nobody seems to care – sorry, there is no “my information” – there is COMPANY INFORMATION.

So, in order to make the MVP -Minimum Viable Product - for effective data architecture, here are some suggestions:

•    Define a common information model to reflect the real company culture; with a model and standards, errors and redundancies fall in at least 25%, which means more reliable data and more effective projects;
•     Define an approach to improve data quality and reliability based on the common model;
•     Define the conversion from the production systems into the common model – there are two possible ways to do it: one is using a data lake based on the common model and another involves using distributed technologies, like edge computing principles and streaming, to produce data faster using the common model and intermediate models/processing areas.
•     Last, but not least, create the decision support area in a data warehouse.

TECHNOLOGY IS APPLIED ON TOP OF THE ARCHITECTURE

And remember, everything that feeds a lake and/or a data warehouse must be improved to generate manageable data sources – even in clouds, packages, ERPs and such. Data is data, no matter what format it takes and must be treated the same way, because it is part of the ENTERPRISE DATA ASSETS. This is what data architecture manages and maintains.

IN SHORT, IT IS MUCH SIMPLER THAN MOST PEOPLE ARE TRYING TO SELL YOU, AND ABOVE ALL, IT IS TECHNOLOGY-AGNOSTIC.

This is true Data Architecture – Happy Sunday!