Why Oracle will continue to win

I was somewhat shocked by the stellar results Oracle recently reported, considering the sorry state of the economy. I even called an analyst friend to find out if maybe there was some house of cards ala Computer Associated that explained the consistent rise in revenue and margin. But I was reminded of two simple facts explaining why Oracle remains dominant:

1. Applications drive database sales
2. Oracle owns pretty much everything

Oracle's acquisition streak has given the company an enormous breadth of offerings (say what you will about quality of the software) and the attempt at offering it's own Linux variant gives it an OS that's passable if not meaningful. But, I don't know that owning the operating system is important to the growth of sales in applications or databases. (Note: Matt Asay wrote a very good post about why Ubuntu should be Oracle's Linux of choice.)

Oracle applications and databases have to run on an operating system, but the operating system doesn't necessarily drive software sales, or sell databases. The OS may be a point of influence, but doesn't drive the dollar values that you get from software.

Meanwhile, Oracle has amassed such a wealth of software that it can not only drive it's own database sales through upgrades and replacements (JD Edwards or Siebel running on DB2 seems unlikely) but it can up-sell databases to customers of BEA or any of the other myriad applications it now owns.

Add MySQL into the equation and Oracle can sell you a database pretty much anytime for any purpose, to support any application (which you can probably buy from them too.)

This leads into some questions regarding Cisco's strategy, based on the idea that hardware should sell applications, as well as IBM's strategy, where services have often sold software and hardware. The future is of course a mix of all of these strategies, but it's not clear that another company is as well positioned as Oracle.

While certainly not unstoppable, Oracle's execution has been very impressive, especially in a down economy.

Follow me on Twitter @daveofdoom.

[cvaldes1831]

From a customer standpoint, here's some of the conversation that probably takes place.

"I need software that runs my company. Who is number one?" Oracle
"What machines do they qual their software on first?" Sun boxes with Solaris
"Okay, I'll take the Oracle RDBMS on a Solaris machine. When can you have this running?"

From a customer perspective, Oracle's acquisition streamlines the tech support since if there's a problem, the Sun support engineer can't point a finger and say, "That's not our problem, that's an Oracle bug" and vice versa. You have one phone number in your Rolodex (not two). You pick up the phone and say "make it work."

[daverosenberg]

Oracle makes it easy to have one throat to choke---even if the rest of the body is taking big money out of your pocket

[cvaldes1831]

Yeah, they say that Oracle's pricing model is grabbing you by the ankles and shaking you upside down; they put you back down when the money stops falling out of your pockets.

Amusingly, Dr. Eric Schmidt bragged that Google got the "world's cheapest Oracle installation." It was when the company was still a startup and he had to convince a lot of people that it was worth it to go for the big iron. While Google is relatively open-source friendly, my guess is that they are still running Oracle on Solaris.

[abcd9009]

@cvaldes1831

Google doesn't use Oracle because neither Oracle nor DB2 and certainly not SQL Server can handle the amount of query requests Google receives every millisecond. Google uses it's own proprietary Google File System (GFS) to handle the queries/user requests. In case you need more details on it - http://en.wikipedia.org/wiki/Google_File_System

[cvaldes1831]

@abcd9009:

Sorry, but I think you are confused. Google uses Oracle for its back-office functions. As you mention, Google uses its own home-grown DBS for its search functions.

[DBConsultant]

Popularity doesn't make it better.
I've been doing DB work since before the relational era, and used all of the RDBMSs since.
I can easily name 3-databases RDBMSs right now that are easier to use, require significant less computer hardware and get added performance and would be a better ROI for any company.
BUT...all that said, companies will still go with "popularity" over functionality.
And, the argument that "they own all the apps" is totally ridiculous. Every system, unless it was owned by Oracle anyhow, has plugs for all popular DB systems, so I don't see how that statement can even be taken seriously.

[cvaldes1831]

IT managers are extremely conservative people. No one gets fired for picking Oracle. If Oracle on Solaris isn't running correctly, you can point the finger elsewhere.

If you pick PostgreSQL running on FreeBSD and it doesn't work, you roll up your sleeves while you enjoy the hot breath on your collar.

The world would be a better place if a few more IT managers took risks, but they're cowards -- and perhaps justifiably so. The older ones have been burned too many times to be adventurous. Or maybe they have families and don't want to be debugging RDBMSes at 2am on a Sunday. A lot of it has to do with the peace of mind.

The Google story is funny since if they decided to install Oracle today, they probably would pay ten times their original amount.

Unless things have drastically changed recently, Oracle buying cycles are pretty drawn out. Much of the revenue they're posting now are probably IT purchase decisions made 12-18 months ago. I would expect to see a dramatic drop in Oracle revenue maybe at the end of this year, reflecting last year's economic contraction.

[supoman]

As an Oracle DBA I both agree and disagree. Sure there are other RDBMS that work just as well as Oracle but I think it's the management tools that separate Oracle from the pack. When you get called at 4:00am and you have to restore
a table or an entire database. You don't want to have to go through a 40 step process. The simplicity of restore database, recover database, recover datafile, flashback table helps me sleep a little easier at night.

[DBConsultant]

There is no RDBMS that I have used that didn't have all the necessary tools to backup/recover a database. All of them are easy, if you have familiarity with them and you've done it a few times.

As for the management tools, it's totally true: oracle does supply quite a number of them. But so does a host of third party vendors and they do it because they feel that the tools oracle supplies are inferior. I know since I used to work for 2-of those vendors. If the oracle management tools were so great, then the third party vendors wouldn't need to exist.

[BradfordStephens]

I think the author may be missing something here. Oracle will continue to grow in certain spaces, probably with RDBMSs, but I think they'll keep losing overall market share. People use the RDBMS as a "Swiss Army Knife" already, treating it as a hodge-podge of key-value store, analytical package, and transactional engine.

In my latest post at http://www.roadtofailure.com , I argue that the end of the Swiss-Army RDBMS is coming -- projects like Hadoop, HBase, and BigTable are going to find a very valuable niche with people who need to scale their data, but don't need everything that SQL can provide. I encourage you to check it out :)