Is Relational the New COBOL? What the History of Technology Tells Us About Change

Adam Hughes

We all know that technology is continuously evolving — otherwise we’d all be riding around in horse-drawn carriages. But what causes one technology to become dominant while another fades away? Are these changes obvious while they’re in progress, or only in retrospect? And what seismic shifts are happening now?

These and other provocative themes featured heavily in a presentation by MongoDB CTO Mark Porter at the recent AWS: Reinvent. Porter’s talk was titled, “Is Relational the New COBOL? What the History of Technology Shows Us About Change.”

COBOL was introduced in 1959, and by the 1970s was the most widely used programming language in the world, powering most mainframe-based software. With the rise of PCs and other advances, COBOL fell from prominence and eventually became a punchline — a stand-in for obsolescence.

The programming language never went away, however. There are an estimated 1 to 2 million active COBOL programmers, and around 220 billion lines of COBOL code still in use, often in mission-critical applications.

But that doesn’t mean COBOL is relevant to innovation. Developers aren’t using COBOL for any new type of development. The language is inefficient, and doesn’t provide nearly the amount of scalability that developers need to build their applications. Porter sees a similar fate for relational databases — still in use for legacy applications, but unfit for innovation and superseded by modern solutions.

The trouble with relational

Much like COBOL, relational databases have a long history. However, as Porter explains, we are long past the point where a relational database is the most productive way to support a new app. Rigid data models and unnatural programming requirements make relational databases far less attractive than modern data platforms, which are enterprise-grade, scalable, flexible, highly intuitive, and run-anywhere.

Here are some of the most interesting takeaways from Porter’s presentation:

Because relational databases are not at the center of new innovation, developers simply aren’t interested in working with them. Porter shared an anecdote about a recent conversation he had with another technology executive.

“He said to me, “Mark, I can’t hire relational people out of school. No one wants to work on relational anymore…the people at my company keep telling me that they will quit if I keep making them work on some of those commercial databases, such as Oracle or SQL Server.”

As the COVID-19 pandemic continues, companies are scrambling to differentiate themselves through innovation. But companies that rely on relational databases are at a disadvantage when it comes to scaling and keeping pace with competitors.

“Enterprises today cannot outsource their innovation. Enterprises during COVID are insourcing their innovation. And when they insource their innovation, they want to move fast. It’s one thing if you can’t scale, it’s another if your competitor beats you to market.”

Will relational really go the way of COBOL — widely used, but only in legacy applications? Porter sees some clues.

“It’s just economics, just like all the technological changes you face in your organization. The articles I researched in 1910 [show that people] thought that cars were this ridiculous thing. They didn’t see it coming. That’s where we are today with relational.”