Data Centers

Will NoSQL be the undoing of Oracle's database reign?

Microsoft and IBM have been building out their NoSQL stories while Oracle largely sits on the sidelines. If they don't embrace NoSQL soon, it could spell problems down the road.

Image: iStockphoto/kentoh

Oracle keeps defying gravity, even as one-time competitor IBM stalls, according to new database popularity rankings from DB-Engines.

And, while Oracle's lead seems all-but-guaranteed for the next few years, NoSQL stalwarts MongoDB and DataStax-sponsored Cassandra are climbing quickly to challenge yesterday's relational database (RDBMS) champions.

Indeed, NoSQL is booming across the database rankings, leading Gartner to acknowledge, "As the operational DBMS market enjoys a dynamic resurgence, new entrants are challenging established leaders."

What's most interesting in all this is how database popularity, broadly measured, compares with Gartner's newly released Operational Database Magic Quadrant. TechRepublic contributor Janakiram MSV has captured five big takeaways from Gartner's report, but here's a sixth:

The database vendors that embrace NoSQL are destined to be the long-term winners.

Falling fast in RDBMS land

Oracle has been the innovation leader in relational databases for decades, thereby earning its spot as the world's most popular database. Oracle also had the foresight to buy big into applications, driving companies to base their database decisions on their application suite (increasingly powered by Oracle).

But, Oracle also managed to pick up the world's most popular open source database, MySQL, when it acquired Sun Microsystems. Small wonder, then, that Oracle has the world's most popular databases, and by a wide margin:

dbranking.jpg
Image: DB-Engines

For everyone else in RDBMS land, however, the last few years have been an exercise in falling fast or slow. IBM DB2, in particular, keeps dwindling, with both Postgres and MongoDB pushing past it over the past two years.

This fading relevance of IBM's database and application business had Oracle's Larry Ellison crowing recently, "We never, ever see IBM."

Apparently few do.

But IBM isn't alone in losing popularity. Microsoft's flagship SQL Server has also been taking some serious dents in the past year, according to DB-Engines. It's not implausible that we'll see MongoDB surpass Microsoft within the next five years.

And while that seems like a long time, remember that RDBMS giants like Oracle have dominated the database market for over three decades. To have some upstart push past them in the space of just 10 years is amazing, and a strong testament to the changing face of data.

Which might just be why Microsoft isn't doomed, after all, and Oracle may need to go shopping again.

Buying a NoSQL clue

We live in a world of big data, a world that doesn't fit neatly into the rows and columns of yesterday's relational databases. Hence, we've seen databases like Postgres attempt to dress themselves up in NoSQL functionality, but the best place to achieve the volume, velocity, and variety of big data remains a native NoSQL database.

This is the opinion of Professor Michael Franklin, one of the industry's foremost database experts, when he calls NoSQL a "complete game-changer":

[W]hat's really fundamentally different about this new-generation data management isn't really isn't just scalability, but it's really flexibility. If you look at the ability to store data first and then impose structure on it later—sometimes this is called schema on read or schema on need—that's a complete game changer.

In response to this shift, Microsoft hasn't been resting on its RDBMS laurels. Yes, it has launched an increasingly popular Microsoft Azure SQL Database that runs in the Azure cloud (now ranked no. 29 in DB-Engines' rankings, down from no. 28 last year), but the real hope for Microsoft is its Azure DocumentDB, a cloud-based NoSQL offering designed to mimic the best of MongoDB.

Microsoft Azure DocumentDB has a long way to go—it is now ranked no. 84 in the DB-Engines ranking. But, last year it was no. 109, and just last month it was no. 92. That's serious progress.

IBM, for its part, has opted to buy NoSQL database bona fides, acquiring document database Cloudant in early 2014. Since that time Cloudant has climbed 13 spots in the DB-Engines rankings, from no. 61 to no. 48.

But given how far behind IBM is, it seems that it should have gone much bigger with its database purchase. Basically, it went bargain shopping and got an also ran. It would have been much smarter to pay more and get a top-10 NoSQL database, because the stakes for IBM are so high.

Microsoft leads the follower pack

Gartner placed Microsoft, not IBM, at the up-and-to-the-rightmost pinnacle of its operational database Magic Quadrant. Despite Microsoft SQL Server's fading popularity, Microsoft scored high marks for thinking differently about data with its Azure DocumentDB, among other things:

Meanwhile, "IBM customers rated it in the bottom third for high transaction rates, and near the bottom for performance. Additionally, IBM scored below the mean for high-speed ingestion and automated data distribution."

And IBM's attempts to resolve these problems are "spread across multiple offerings," which is tantamount to telling enterprises they might as well buy best-of-breed from different vendors than attempt to one-stop-shop with IBM, where none of its database offerings are best in class.

Lastly there's Oracle, the company that leads almost despite itself. After all, as Gartner finds, "A growing number of users of Gartner's client inquiry service express dissatisfaction with Oracle's 'draconian' pricing and auditing policies; the number of these users looking for alternatives to Oracle software is increasing."

That number, even without the pain of doing business with Oracle, will increase as big data drives enterprises to get a rich NoSQL experience, something that Oracle doesn't offer.

Over time, Oracle will likely buy a leading NoSQL vendor, with DataStax-sponsored Apache Cassandra the likely choice. This would be a smart move, because with Oracle's continued earnings misses, it's running out of time before it has to catch up to the future of data and start selling what its customers increasingly want: Big data-friendly NoSQL.

Microsoft got the memo and is building aggressively. IBM got the memo and has tried to go cheap. Oracle needs to read the memo and go big.

Also see

About

Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.

Editor's Picks

Free Newsletters, In your Inbox