Don't read this, Oracle... It's the rise of the open-source data strategies

Could this be the end of RDBMS?

Oracle acrobatics in the cloud

Comment Oracle is industry’s single largest database vendor – which was great during the days before cloud and open source.

Now, however, the on-prem RDBMS giant faces a challenge, one which is compounded by cloud. According to Gartner, Oracle owns just a tiny sliver of the cloud infrastructure market – 0.3 per cent. But while that tiny figure doesn't seem to account for how cloud hurts Oracle's database business, there's something else to consider: a developer's first decision is what cloud platform they'll use.

The database a developer elects to use then follows from the options made available by that cloud platform. By making it cost-prohibitive to run Oracle DB on the public clouds that developers actually use – AWS, Microsoft Azure, and Google Cloud – Oracle has tied its database future to the mast of its sinking cloud ship. Who reaps the benefits of its myopic strategy? Open-source databases.

Opening up the database

Yes, open-source databases. While databases like MySQL, PostgreSQL, MongoDB, and Apache Cassandra have long scored points with web developers, historically they didn’t compete on Oracle’s core database turf.

But that was then, this is now. According to recent Gartner analysis, open-source databases now constitute 7.6 per cent ($2.6bn) of the global database market, worth $34bn. If that doesn’t seem like much, consider that over the past two years the open source DBMS market averaged 75 per cent growth, compared to a more tepid 7.7 per cent growth in the total market.

Where is that growth coming from? In part, it reflects developers’ desires to run new applications with modern databases. Those decisions have been made much easier by AWS, in particular, which has taken many of the most popular open source databases and turned them into services, removing the complexity of managing them. As such, according to DB-Engines, which ranks database popularity across a number of factors, half of the world’s most popular databases are now open source.

Widen that aperture a bit and you get MariaDB, a fork of MySQL, in the 13th spot. Widen it even further, and you see a host of other open source databases, and particularly those that AWS has turned into cloud services, roaring up the popularity charts.

As noted, developers are increasingly turning to the cloud for their databases, spurring a host of database services from Amazon, Microsoft, and Google to rocket up DB-Engines’ rankings. For those who want the option of running a database in the cloud or in their datacenter, the primary choice is between PostgreSQL and MongoDB.

The former is often the preferred choice for a developer who wants to stick with a relational database but wants to remove the cost or complexity of Oracle. MariaDB is also increasingly an option for this crowd. MySQL, somewhat out of favor since its acquisition by Oracle, has been declining in popularity over the last several years.

The other option, MongoDB, gets picked when a developer is refactoring her application and needs a significant boost in developer productivity or the scale-out architecture that MongoDB’s document database affords. Either way, the choice is for open source, not Oracle.

It would be comforting to the Oracle faithful to believe that this open source onslaught isn’t having an impact on the database behemoth. It would also likely be wrong. In a conversation with Gartner analyst Merv Adrian, he noted that Oracle has lost market share every year since 2013. As a club, the top-five database vendors have seen their collective share of the market drop from 91 per cent to 86.9 per cent since 2011.

While open-source databases can’t take all the credit for that slide, what with proprietary cloud databases like Amazon DynamoDB or Microsoft CosmosDB also in play, it’s almost certain that open source databases are wreaking havoc to the tune of billions of of dollars.

Nor do those dollars tell the whole story.

Gartner, after all, measures market share based on revenue. Open source databases, however, are used for free far more often than they are used "for fee". Recognizing this fact, Gartner posits that “a good overall rule of thumb is that paid customers only account for 1 per cent to 5 per cent of the actual user base.” In other words, open source databases may earn $2.6bn for their vendors, but they annihilate orders of magnitude more paid value that other vendors like Oracle might otherwise earn.

But wait! It gets worse for Oracle.

Making developers happy

The biggest risk to Oracle is not the cost of an open-source database like MongoDB or PostgreSQL. While license fee savings from open source hit 100 per cent, open-source databases also yield significant savings in hardware costs. All in, companies can expect to save 70 per cent by shifting from Oracle to a database like MongoDB (even once you account for the cost of migration, re-skilling DBAs, etc.) On the AWS platform, the list price for running Oracle (RDS) is $25.68 per hour. Running PostgreSQL or MySQL (RDS) is 1/8th to 1/10th that cost.

As big as those savings are, however, the bigger cost differential derives from developer and DBA productivity.

For DBAs skilled with Oracle’s database, they can often manage up to 25 database servers, on average. That same DBA can manage a million database servers on Amazon’s RDS, thanks to the benefits of automation. Talk about scale.

On the developer side, given that developers are the new kingmakers, as Redmonk likes to remind us, shifting to an open-source DBMS is much more about super-charging developers than any belt-tightening exercise around license or hardware costs. Mat Keep, director of product marketing at MongoDB, put this in a personal context:

When I joined MongoDB, about 5 per cent of all projects were relational migrations - now it’s 30 per cent as companies look to transform. Cost can be a factor, but more often it’s development speed and running at scale. It’s not unusual to see developer productivity up 3 to 5x after switching [from an RDBMS], coupling MongoDB with a shift to cloud, microservices, and agile/devops.

Even in areas where Oracle likes to trumpet the richness of functionality it offers, like Oracle HA, the reality is that much of the “richness” is actually external to the database itself. You “have to add a ton of stuff outside of the database [to make it work] for replication, failover, monitoring, etc,” Keep says. Of course, this being Oracle, each of these add-ons is sold separately, resulting in a fat price tag and a seriously complex system to manage. Even worse for Oracle, the only way for a developer to get access to such Oracle extras is on the Oracle cloud, which basically no one wants to use.

Nor are such add-ons all that revolutionary. On its latest earnings call, Oracle said, with a straight face, “[T]he amazing thing about the autonomous database is it's the only database in the planet that requires no human labor to administer the database.”

This is 100 per cent false. Behind the scenes Oracle has scads of people running around to keep the lights on, even as AWS and other cloud vendors deliver truly automated databases at dramatically bigger scale. If anything, Oracle is way behind the cloud versions of open source databases.

But, but, but…

Of course, there’s a reason Oracle continues to makes bajillions of cash selling its database: it has decades of experience building databases, and does so very well. The problem, however, is that what worked for Oracle 10 or 20 years ago is looking less and less ideal today. Its strengths like serious scale-up architecture are now relics of a bygone era.

Name the last big company to stake its future on Oracle. Salesforce probably comes to mind, though rumors have rumbled that it’s not happy with its choice. Meanwhile, the other big SaaS companies like Workday have been building on open-source databases like MySQL, often running them on AWS or other clouds. As enterprises move to distributed computing, they’re trying to minimize the cost of failure, separating, for example, storage from compute. As they do so, Oracle simply isn’t a consideration.

Open-source databases like PostgreSQL and MariaDB remove the bureaucracy inherent in acquiring Oracle’s database. PostgreSQL, in particular, has made it relatively straightforward to migrate things like stored procedures from Oracle to PostgreSQL.

Even non-relational databases like MongoDB have obviated the need to stick with Oracle. Oracle has conditioned developers and DBAs to think such relics of relational data modeling – like multi-record transactions – are essential. They’re not, and more companies are discovering that they either don’t need the “full-fat” Oracle database and can go with a lower-cost relational open-source database, or they’re seeing that they don’t need a relational database at all and can instead benefit from increased developer productivity and improved scale that a NoSQL database (with ACID guarantees) like MongoDB offers.

What does this mean to Oracle? Given how much friction is involved in moving to alternative databases, Oracle’s database should last a long, long time. Much more likely is the fact that we’ll see all the growth go to open-source databases and cloud databases (with particular growth in those databases that are both open and cloudy). Expect, also, Oracle to keep buying SaaS companies as its future becomes fueled more by SaaS applications and much less so by database dominance. ®

Matt Asay is Head of Developer Ecosystem at Adobe.




Biting the hand that feeds IT © 1998–2018