I've been running a rather large website with MySQL for the past fifteen years. There was a period when I regretted that choice and used something else. Today I

I've been running a rather large website with MySQL for the past fifteen years. There was a period when I regretted that choice and used something else. Today I'm using MariaDB and the TokuDB storage engine, and I'm so thankful that I never migrated to Postgres. Like many people I investigated the NoSQL movement for potential applicability, and almost swallowed the hype. As I investigated more, I realised: 1. There are some specific instances where a NoSQL engine makes good sense. They're a valid option and should be considered depending on the application. In my experience though, well formed RDBMS structures are the better option in the vast majority of applications. 2. Most of the hype and growth came from people who (a) were using the abomination known as ORMs which are the canonical example of a round peg in a square hole; and/or (b) didn't know how to build performant RDBMS schemas. For these people, the NoSQL engine was fast because it was the first engine they actually learned how to optimise correctly.