Home World Technology Can’t kick the relational database habit? Serverless computing may help

Can’t kick the relational database habit? Serverless computing may help

1
0
SHARE

Relational databases never say die. Even, of course, when they probably should. If anything, as DB-Engines database popularity rankings show, the stodgy relational database management system (RDBMS) remains top of the database charts, and has been growing stronger even as applications seem to fit into their rows and columns less and less.

Partly this is a testament to the RDBMS evolving to meet new needs, but mostly, according to Movivo CTO Paul Johnston, it’s an indictment of our own unwillingness to do so.

The little RDBMS makes way for big data

We really shouldn’t be having this conversation. The three V’s of big data (variety, velocity, and volume) all demand a very different database than traditional RDBMSes can supply. That’s the theory, and it’s one that database experts like Michael Franklin have hailed, calling NoSQL databases “a complete game changer” for big data.

SEE: Will NoSQL be the undoing of Oracle’s database reign? (TechRepublic)

And yet…the market isn’t going gaga for NoSQL. Or, rather, it’s not doing so at the expense of RDBMS, as DB-Engines shows:

screen-shot-2017-07-06-at-1-24-07-pm.png

DB-Engines.com

MongoDB had been climbing for years, ultimately displacing PostgreSQL for the #4 spot—only to lose it to a resurgent PostgreSQL. This, despite MongoDB gaining converts on a regular basis.

Old habits die hard

More about Big Data

Johnston, at the forefront of the serverless trend, has fond memories of relational databases: “I still love RDBMS. But I love RDBMS like you love an old dog that can faithfully go on a walk with you, but no longer does the tricks that younger dogs can.” The problem, he wrote, is that RDBMS is generally the wrong way to achieve scale: “As a developer building a solution, one of the things you have to now consider is data storage at scale. And unfortunately, that almost certainly means not using RDBMS by default.”

Sure, you can get an RDBMS like MySQL to scale. As Linux Foundation executive Chris Aniszczyk told me, “It actually can scale” quite well. Evidence for this abounds, as Facebook, Google, and others have pushed MySQL and other RDBMSes to new limits.

SEE: NoSQL keeps rising, but relational databases still dominate big data (TechRepublic)

But RDBMS isn’t the default for delivering scale, Johnston stressed. It remains the industry’s default, however, not because it’s the best solution, but “because [it’s] a safe place for a developer to go.”

This was fine when applications offered leeway for RDBMS complexity. In a world of serverless, “where a lot of short-lived connections are made over and over again,” an RDBMS yields “either the overhead of connecting and disconnecting (slow), or you end up with the overhead of connections not being dropped and the server reaching its maximum connection limit.” No bueno. Data storage and retrieval in this context become painful with an RDBMS, but natural for a NoSQL data store like DynamoDB.

As Serverless becomes more the norm, we should expect to see the database industry follow. Or, as has been the case for the last few years, will we see more developers fitting a square RDBMS peg into a round serverless hole? Time will tell.

Also see

dbmanagement.jpg

dbmanagement.jpg

Image: iStockphoto/denizbayram

LEAVE A REPLY

Please enter your comment!
Please enter your name here