

It’s like trying to untangle the Gordian Knot in software. You may not even be able to change important database characteristics if your team doesn’t have ownership of it. As your scale increases, your options become even more limited. And each query can only be optimized so much without requiring pervasive schema changes. Only so many queries can run concurrently in the same place. In some sense, relying on a centralized database can feel deceptively comforting because you only have to worry about one thing.īut it is undoubtedly a natural pinch point. The problem at the heart of all this is that the database has been cemented into every area of your company-even those that would be better off without it. The list of negative consequences goes on. Good ideas go unexecuted, because no one wants to be the one to disturb something so temperamental. Engineering decisions come to a crawl, since every schema change has the potential to bog down multiple teams in endless change board review meetings. Queries become slow, taxing an overburdened execution engine. When a company becomes overreliant on a centralized database, a world of bad things start to happen.
