Back in November of 2005, when polled on application areas that they would be reluctant to trust to open source, the following group of individuals – nearly to a person – pointed to mission critical databases:
- Gary J. Beach (Editor, CIO Magazine)
- Ameet Patel (Former CTO, LabMorgan (JP Morgan Chase))
- Don Haile (President & CIO, Fidelity Investments)
- Ron Rose (CIO, Priceline.com)
- Jamie Cash (CIO, NLG)
- Barry Strasnick (CIO, CitiStreet)
What’s changed since November of 2005? Little.
With the rare exception, most of the open source database vendors have declined to storm the castle, believing – correctly – that they were not in a position to compete effectively in such markets. Many also accurately anticipated that even if they were technically equivalent, the strategic importance of the software and the attendant risk to changes to it would preclude migrations, wholesale or otherwise.
At which point, open source database vendors had a choice: pack up the tents and head home, or attempt to capture or create new and emerging markets. Some undoubtedly chose the former, but several pursued – successfully – the latter course of action. Unable to duplicate their proprietary rivals in profit margins and revenue, they’ve instead eclipsed them in adoption.
All of which explains why I’m surprised that anyone is surprised by the 451 Group’s report on the impact of open source on the database market. The word choice is controversial and surprisingly harsh (Matt calls it a glass half empty) from the perspective of open source database vendors, but I find little to argue with in its assessment of current market conditions:
One of the key findings is that open source software has had a superficial impact on the enterprise database market in that adoption has been widespread but shallow. While open source databases have been widely deployed for Web-tier applications, there has been minimal adoption in the enterprise application tier, and adoption for enterprise applications is at this time limited to certain specific application workloads.
While I believe the above fails to acknowledge the degree to which adoption is widespread, the landscape painted is otherwise accurate, in my opinion. The impact of the open source vendors on their proprietary competitors has been undeniable – unless you think IBM and Oracle would have released free if limited versions of their products in the absence of such competition – but probably negligible from a financial standpoint.
From those that would dispute the conclusions of the report – which, it should be noted, I don’t have access to – Savio, the open source myth buster, would take away the counter-arguments that a.) revenue measurements are apples to oranges and b.) the OSS market is still young. With an argument that I only partially accept.
But while I cannot dispute the assertions regarding the current state, it seems likely that I’d draw significantly different conclusions from it than Matthew Aslett or Savio. Primarily because while I value both revenue and volume, I tend to bias towards the latter when considering long term projections, given that it’s built the largest software business on the planet.
The 451 Group’s report concludes that the enterprise applications market has been largely impenetrable to open source databases to date, and I would agree. But what is more plausible over the longer term: that EnterpriseDB and MySQL gain the technical competence to become at least adequate in the applications market? Or that DB2 or Oracle will be able to downsize for web tier deployments while preserving their margins?
If IBM’s investment in EnterpriseDB is any indication, it would seem that they’d argue the former. And I would agree with that as well.
If we’re to consider revenue as an guideline metric, as Savio requests, we must also consider opportunities ahead for revenue growth. Give closed source vendors a massive edge in the former, but then ask yourself whether open source vendors have equally sizable advantages in the latter.
Depending on how you answer that, you’ll know whether you consider the open source database glass half full, or half empty. You’ll also know whether you believe the impact of open source vendors will continue to be superficial.
Disclosure: EnerpriseDB, IBM, and MySQL/Sun are RedMonk customers, while Oracle is not.
Matthew Aslett says:
March 28, 2008 at 6:11 am
Thanks for the comment Stephen. So far the reaction to the report has centered on that one (admittedly key) finding that adoption has been widespread but shallow, but there is clearly more to it than that, and we do see and mention opportunities ahead for both adoption and revenue growth. As I wrote the other day: “The adoption of open source software for non-mission-critical applications and new projects will continue, and we expect to see open source databases gradually surround proprietary database deployments. At this stage, mainstream customers will begin to reevaluate their core database management offerings and examine whether open source is a viable option for mission-critical applications.”
Mike Dolan says:
March 28, 2008 at 6:44 am
Is it the database that’s important or the interfaces through which you access and manage data? When looking at shallow/deep, the answer to that question could make a big difference in the conclusion about where these databases will/can play. Yes, I’m being intentionally vague 😉
Niraj J says:
March 28, 2008 at 1:25 pm
On “But what is more plausible over the longer term”.
The Enterprise penetration of open source DB is a function of the success web startup’s and their speed to transform themselves into large enterprises.(typically revenue > 1 billion)
As you correctly pointed out , there is no point of competing in the Enterprise DB space (As defined today) . my guess is that the current Enterprise DB space will evolve and split into specialized Databases like (GFS , Hadoop based DB) , HP Neoview , CouchDB , simple DB , commodity RDBMS(my SQL) .
I would think that thought leaders for Enterprise DB vendors would have already started to look at their RDBMS business as cash cow rather than a growth business.
Anne Truitt Zelenka » links for 2008-03-28 says:
March 28, 2008 at 4:55 pm
[…] tecosystems » Open Source Databases: Shallow But Widespread (tags: open-source databases enterprise web revenue volume) This entry was written by Anne Z and posted on March 28, 2008 at 5:54 pm and filed under Delicious Links. Bookmark the permalink. Follow any comments here with the RSS feed for this post. Post a comment or leave a trackback: Trackback URL. « links for 2008-03-24 […]
Mike Johnson says:
March 30, 2008 at 9:13 am
Your analysis of the database market is right that there’s a split between “enterprise” databases and the web side, but I think your idea of future inroads is a little off.
Applications are moving swiftly to being entirely web based in large corporations. Easier to deploy, easier to write, easier to maintain, and no install. It means you don’t have to talk to the desktop guys or change 1000 desktops to deploy, just throw up a web server. If you’re running open source for that already, the proprietary databases are marginalized to running backend, legacy apps.
Sooner or later somebody’s got to ask “why are we spending $1 million a year in Oracle fees when most of our stuff is running on these free open source databases anyhow?”
My company was bought out by a bank and they’ve been (somewhat unwillingly) running several Postgres databases for around 5 years now. Never had a problem with them, they just run.
As time goes by, the argument from the database support team that they don’t and will not support Postgres grows weaker. It starts to look like a refusal to learn something new rather than a sound technical argument. But it won’t be up to them forever.
March 31, 2008 at 2:52 pm
“As time goes by, the argument from the database support team that they don’t and will not support Postgres grows weaker. It starts to look like a refusal to learn something new rather than a sound technical argument. But it won’t be up to them forever.”
Couldn’t have stated that better myself. We’ve been using the open source SAPDB on a Linux x64 platform in a mission critical role for years now. What you don’t hear is that the server never needs reboots (thanks Linux) and the db just runs and runs without complaint. It can handle loads far greater than our sister db MS-SQL v7 which constantly needs attention. The load on the MS-SQL server is far less and yet needs more resources and attention. OSS-DBs aren’t ready for mission critical … PA-Leeeeez!
Henry Keultjes says:
March 31, 2008 at 5:03 pm
http://www.openqm.org is functionally identical to IBM’s UniVerse and UniData and similar proprietary enterprise databases that together have millions of seats.
Mansfield Ohio USA
March 31, 2008 at 7:55 pm
When OSS ERP systems are more regularly deployed with an OSS DB as a back end. That will systematically increase enterprise OSS DB usage.
Right now I think that when you buy an ERP/CRM system they recommend Oracle/MS SQL and DB2. When you look at OSS solutions MySQL is the standard thing to use and Postgres can easily be used to replace MySQL if so desired.