Red Hat CEO suggests Oracle is feeling the heat

Rob writes "The previously rosy relationship between Oracle Corp and Red Hat Inc appears to have soured following Red Hat's acquisition of JBoss Inc and Oracle CEO Larry Ellison's suggestion that his company could move into the Linux business. Red Hat's chief executive, Matthew Szulik, has written in response to a recent interview with Ellison in which Ellison suggested the company would be interested in distributing and supporting Linux. "Is it possible that the dominant provider of databases feels pressure from its long-time partner, Red Hat, because of our recent purchase of an open source middleware company, JBoss?" Szulik asked, although he also played down suggestions of a "showdown" between the two companies."

Comments

Link to the actual letter.

[Whiney Mac Fanboy]

If anyone feels like reading the actual letter Szulik sent to the Financial Times, they can read it here.

There is also a no-reg-required mirror at the zimbabwe open source software society.

The most intersting part of the letter is where szulik puts a new twist on the (always perfect) car / computer analogy

I have a much better appreciation of the challenges the Japanese carmakers faced when attempting to break into the domestic US market while competing against historical industry practices and the personal networks that stood in the way of customers having access to a lower-cost, higher-value alternative. Open source software and Red Hat continue to face similar challenges. But in the end, through innovation and a commitment to the customer, the Japanese automakers delivered choice to the customer. The US automotive industry is a good case study, in comparison to the state of the domestic US software industry.

Well put.

Showdown? lol

[Toby The Economist]

Oracle could swallow Redhat without even needing a context switch.

Re:Showdown? lol

[Jason Earl]

There's a funny thing about market economies. The market produces what people have shown an interest in paying money for. If Oracle were to buy Red Hat with the intention of shutting Red Hat down, then you can basically guarantee that several other "Enterprise" Linux distributions would spring up as if by magic. This is especially true if Oracle paid current market prices for Red Hat. The source code in question would still be available, and there would be a large community looking for a new home.

In the long run Oracle is likely to have the same problems as Sun. Like Sun, Oracle's real problem is that Red Hat offers a software stack that is competitive with Oracle's software stack while maintaining an R&D budget that is a couple orders of magnitude smaller than Oracle's budget. Oracle's size is precisely the problem. As commodity software becomes more and more widespread the ridiculous profit margins that Oracle needs to survive will get harder and harder to produce. Sure, there are lots of Oracle customers that can't really afford to move to a lower cost but less featureful software stack, but Oracle is going to find that an increasing number of its customers are unwilling to pay for features that they don't really need or use. Lots of technical folks get all excited about "Enterprise" software, but in the long run inexpensive commodity software that actually gets used tends to move up the technology stack and crowd out software that relies on the huge profit margins that can be found at the high end of the spectrum. Red Hat's cost structure is designed around taking advantage of the much lower profit margins associated with commodity Free Software. Oracle's cost structure, on the other hand, is designed around the much higher profit margins that Oracle has historically been able to squeeze out of the market. Oracle can pretend that it can compete with Red Hat, but really it can't, not without shedding a lot of its workforce. If Oracle were to lower its workforce so that it was competitive with Red Hat then customers that are currently paying huge margins for Oracle products and services would undoubtedly take their business elsewhere.

Everything old is new again

[vallee]

The thing is, this affair of Oracle considering entering the Linux support arena and even shipping its own Linux distro is not new. Not even close.

It dates from 1998, during the initial launch of Oracle 8i. Since then, and arguably for even longer, Oracle has had a consistent strategy of undermining the role of the operating system by taking on more and more of the critical duties into its own code base. Linux plays into this strategy marvelously well. Except, here's the rub. Redhat is not interested in the furtherance of this agenda. Redhat wants the operating system to remain a key part of the enterprise IT infrastructure.

I wrote an interesting article on my blog titled "Oracle & Linux, Ancient History" on this subject last week, and the article links to the web archive of my original post about Oracle and Linux and Oracle's strategy to undermine the OS from 1998. The original article's title was "Why Oracle 8i Will Remodel the OS Landscape" and ultimately what we're seeing now in the tension between Oracle and Redhat is the materialization of Oracle's vision of the operating systems' role chafing on its longstanding partner.

Cheers,
Paul
P.S. Pythian DBAs post on our group blog at http://www.pythian.com/blogs/.

Good idea for Oracle

[Bert64]

It would make a lot of sense for Oracle to produce a complete dedicated package that didn't require an OS already be installed. Most Oracle database systems are dedicated machines anyway, so having the entire package supported by a single vendor instead does make a lot of sense. No more Database vendor blaming the OS vendor :)

Re:Good idea for Oracle

[WinterSolstice]

That'd be nice, except that with IBM DB2 on IBM AIX running on IBM RS/6000 systems, they still point fingers. I have repeatedly been forced (as a customer) to make different divisions within IBM talk to each other. It would seem to be good, but it really is almost the same.

-WS

Re:Good idea for Oracle

[DavidpFitz]

It would make a lot of sense for Oracle to produce a complete dedicated package that didn't require an OS already be installed. Most Oracle database systems are dedicated machines anyway, so having the entire package supported by a single vendor instead does make a lot of sense. No more Database vendor blaming the OS vendor :)

Funnily enough, Oracle did this with the Oracle Appliance for 8i and 9i. They never made an Oracle 10g appliance. It was exactly as you outline above - it's a preconfigured software stack to run Oracle. Everything already set up in a basic form. Made life very easy when needing to test a DB agnostic app against Oracle.

Oracle have since removed reference to it from their site.

Re:mysql?

[fireboy1919]

Umm...nobody who uses databases ever claims B). On the constrary: MySQL has less features than most of the database engines out there.

However, nobody cares about most of the extra features. So let's change B) to: supports all of the features of Oracle that most people care about

Did you know that Oracle comes with something to do a text search on almost any document type, including those accessible through URLs? And that you can do fuzzy searches based on that, and that the database can learn to give better results via an expert system? It's a pretty nice search engine. Does MySQL come with a search engine?

Also, if I've said it once, I've said it a million times: don't exaggerate. The personal use version is free, and Oracle is $5k per copy for the one-processor, coarse-grained security model. The high-end one for clustering that you seem to be thinking of is $40k, not $200k.

My guess is that the market share of stupid people who buy Oracle when all they need is MySQL is dying. However, there really are people who want to do extremely sophisticated stuff that only Oracle is providing. Oracle's real up-and-coming competition for their real market is Google, I think.

If Google will do all the indexing and does a better job of managing your data without you having to even configure it, then why should you manage it with Oracle?