Were relational database management systems first used for transaction processing?

Martyn Richard Jones, Madrid, 3rd October 2023

Narrator: Relational database theory is over fifty years old. Relational database technologies go back more than four decades. Some claim that relational databases were first used for online transaction processing applications. Say what?

Dud: Pete, do you think relational databases were first used for transaction processing applications?

Pete: No, Dud.

Dud: And?

Pete: If anyone claims that relational database systems were first used with operational applications, treat everything they say cautiously.

Dud: Why do you say that, Pete? Are you being contrarian again?

Pete: Listen up, Dud! The first relational database management system products were used for reporting needs, such as reporting on data in databases designed using dimensional modelling. There was a good reason for this, as none of the implementations even came with a usable audit trail facility. So, it wasn’t at all OLTP-friendly at the beginning. Businesses, mainly banks, use them to produce reports and take the production of reports from business-critical mainframe platforms to cheaper and less reliable platforms using Unix or proprietary operating systems.

Dud: I didn’t know that, Pete.

Pete: What’s more, Pete. Even today, many of the largest businesses don’t trust their mission-critical operational systems to use relational database technology.

Dud: Fascinating. I’d never looked at it this way before. So, in the beginning, it was pretty different.

Pete: these relational engines were absolutely useless for transaction processing of any importance. They lacked even the most rudimentary elements required for transaction processing. They excelled in producing reports if the report producer properly understood relational algebra, first-order logic, set theory and when not to use indexing. Still, they were absolutely unsuitable for anything else, Dud.

Dud: But people say relational databases were first used to manage large amounts of transactional data. Are they right, or are they right?

Pete: They are talking absolute nonsense, Dud. Many of the original uses of relational databases involved much smaller amounts of aggregated data. Far less data than you would have found in operational mainframe systems.

Dud: So, why do the experts say otherwise?

Pete: It’s easy, Dud. They don’t know, so they make stuff up, hoping they will get away with it. In the long run, it doesn’t matter. In the long run, we are all dead. But, in the here and now, they are irritating little scamps who do nothing more than bring my profession into disrepute. Although, to be fair, some were fooled into thinking that Codd and Date’s description of relational databases was an actual implementation. Which of course, it wasn’t.