April 1, 2018 - Dale Kim | Big Data Ecosystem

Hadoop Is Not Dead, It Just Has No Life

It’s cool today to proclaim that Hadoop is dead. It’s perhaps even cooler to ignore Hadoop altogether. Just ask some of the cool technology vendors today like Cloudera, Hortonworks, and MapR, who no longer find the need to aggressively tout their Hadoop roots. And where are Hadoop World and Hadoop Summit, the main conferences for Hadoop? Not gone, just cleansed of the stench that encircled their beloved eponymous platform. But is this the start of the death of Hadoop?

This is quite an intriguing phenomenon. Just as Hadoop was becoming a household name after 10 years of existence (while creating new terms like “BI on Hadoop”), people are now running from it as if it were an object-oriented database (though admittedly that’s a bit too harsh to say, because Hadoop at least had happy users). We’re not talking “trough of disillusionment,” it’s more like the “trough of disembowelment.”

So, what happened? The fact is, the name didn’t have longevity and people turned on it. Don’t get me wrong, Hadoop is a great name for a stuffed elephant (I mean a toy one, of course), but clearly less appealing for the foundation of your modern data architecture. It didn’t help that the creator of Hadoop pronounced it HAD-doop, while you pronounced it huh-DOOP.

Is the thing that we once proudly called Hadoop going away any time soon? Of course not. Businesses are still buying the-technology-formerly-known-as-Hadoop in droves, because we all know it actually works well. And if it’s to continue proliferating, we can’t keep ignoring the name, nor can we call it TTFKAH. Or can we? (I was told we can’t.)

Fortunately, we at Arcadia Data have adopted a new name for the thing you get when you work with the-technology-formerly-known-as-Hadoop vendors. Since the key component of Hadoop is the file system, we thought to put the emphasis of the new name there. We call it the Distributed Universal Filesystem for Universal Systems, or Dufus. Some say the name is redundant, but I believe it has a nice ring to it. I see a “Strata + Dufus World” in the near future.

Hadoop is not dead

Dufus wasn’t my first choice. I wanted to follow Doug Cutting’s example and use the name of my son’s favorite stuffed toy which he got as a free gift from a Las Vegas casino. I want to say it’s a stuffed monster, but it actually looks like a big orange gumdrop with two arms and the word “Boo” written all over it. We call him “Orange Boo” and I was hoping to one day talk about “Arcadia Enterprise for your Orange Boo clusters,” but that didn’t work out.

We hope the open source community follows our lead, because “Apache Dufus” sounds like a winner to me. This should restore the greatness and respect that was always well-deserved by Hadoop.

And while I’m here, something else has been bothering me for a while. Ever notice that NOBODY likes the term NoSQL? Whether you pronounce it as no-SEE-kwuhl or no-ESS-cue-EL, the name is an abomination. Never mind that SQL is everywhere, so a term that contradicts that is going to have problems. Maybe it’s time to change it. At least one company has changed the name of their offering to a “wedding database” or something weird like that, but they’re going to have to try harder than that.

The name shouldn’t try to emphasize a difference from SQL-based databases, especially since all NoSQL databases support SQL. If relational databases are named for relations (you knew that a “relation” is another name for a table, right?), then maybe NoSQL databases should be named for the chaotic mess that they store in each record. Chaotical databases? Maybe the brainstorming is best left to the chaotical database vendors.

And also, what about containers? Should we really use the same term for application virtualization as we use for Tupperware? Will containers face the same ostracism as Hadoop, simply because of the mishandling of the branding?

The bottom line is, if some of these great technologies are going to survive, they need to have better names. Like Skynet or HAL. Or maybe something less doomsday-ish like Data Blaster. I gave you the new name for Hadoop; now help us all out and find new names for NoSQL and containers.

Related Posts