Federated Data Governance – A Fireside Chat

What are your thoughts on centralized versus federated data governance? To get to a sort of a hybrid governance model where you are acknowledging the fact that the data could live anywhere, in any format, and your ability to essentially create a logical version where you can connect the sources in a meaningful way and provide governance on top of it. It's meaningful.

If you give responsibility of governance to individual systems, individual systems only know what they are responsible for. If you go with a completely centralized model, when centralized governance model don't necessarily have detailed understanding of how the governance is maintained by individual systems. So it's very difficult to take hundreds and thousands of systems, decentralized governance systems, and create a centralized big system.

So the future seems to be somewhere in the middle. But centralized governance for setting policy and then localized for the domain expertise. It's kind of how countries are run.

Most democracies have federal and state. The federal government does something. State government does something because there are certain problems that are common to all and there are certain problems that are unique to those individual systems.

And there's no one approach, there's no one size fit all kind of a thing where you can go and say, well, everyone should do this. Yeah. I think because data teams that are defining governance policies, right, and that are enforcing governance policies tend to be small when they're centralized and they don't have that domain expertise.

So now they have to become domain experts on every part of the business so that they can determine.

Exactly and imagine, you know, going through acquisitions, you acquired a company and they have a different system. So unless you are outcome driven saying we want you to eventually accomplish this, and you are flexible on how someone can get there, I think it's very difficult to run business that way.