Should Oracle's Linux strategy be...Ubuntu?
Oracle has gone on a buying spree in the past few years, consolidating an impressive portfolio of market-leading technology. But there's one thing it still lacks, despite awkward efforts to fill the void: an operating system. Though Oracle has unsuccessfully courted Red Hat as an acquisition target for years, its affections might be better placed on Ubuntu.
Yes, by acquiring Sun, Oracle is gaining Solaris, but as Red Hat CEO Jim Whitehurst indicated in the Red Hat earnings call on Wednesday, the exodus of Solaris-to-Linux users continues apace, as Sun's attempt to neutralize Linux's appeal with OpenSolaris have had zero effect on stopping the exodus.
Oracle Enterprise Linux (OEL), a Red Hat Enterprise Linux (RHEL) clone, has failed to dent Red Hat's dominance, with Red Hat renewing all of its top 25 deals (again, and at 120 percent of renewal value) that were up for renewal, losing none to OEL. As such, RHEL's dominance remains a festering sore for Oracle's ambition to own a complete enterprise software stack. So long as Red Hat owns the foundation of that stack, it remains a real threat to Oracle.
So long as it's easier for Oracle's sales force to sell Oracle applications and databases on RHEL than on OEL, OEL will continue to bluster but fail.
Oracle could buy Red Hat, but with Red Hat earnings consistently strong, Red Hat is arguably too pricey to be a viable takeover target, as The Register opines. Besides, Red Hat has shown no desire to jump into the arms of its Redwood Shores business partner and competitor.
All of which makes me think that a strong partnership with Canonical for Ubuntu, rather than continuing to feed Red Hat, could be the right Linux strategy for Oracle.
Ubuntu is the clear community choice in Linux distributions, dominating Linux "desktop" adoption and also claiming a solid foothold on enterprise servers. Unlike OEL or Novell's Suse, Ubuntu comes with built-in enterprise momentum, albeit still at the grassroots level. Oracle's sales force could sell Ubuntu as a complement to Oracle technology, unlike OEL which is a difficult sales pitch.
The spirit to sell OEL is willing, but the flesh is weak.
As just one data point on this enterprise readiness to accept Ubuntu, my company, Alfresco, an open-source content management vendor, has seen adoption of Ubuntu rise to 37 percent of all trials of our Enterprise product, versus 28 percent for RHEL and Fedora.
A year ago, Ubuntu was making serious headway, but today the interest seems to be migrating to using Ubuntu for real enterprise deployments.
Other open-source companies I advise are seeing similar Ubuntu traction, a fact that is perhaps not lost on Dell and other hardware manufacturers that increasingly preload Ubuntu on their servers, Netbooks, and laptops.
Ubuntu, in short, has community momentum. What it lacks is the blessing of a major software vendor. Oracle, with its heft, is a kingmaker, and could give Ubuntu the "enterprise-ready" branding and certification that still elude it.
Not coincidentally, it was Oracle, more than any other company, that blessed Red Hat as the default enterprise Linux distribution years ago. But for Oracle's support for Red Hat Enterprise Linux, we'd almost certainly have a very different Linux market today, one where Novell's SUSE and other Linux distributions would have more respectable market shares compared to RHEL.
Oracle has the ability to make Ubuntu a mainstream enterprise server player. The question is whether it wants to.
As Matt Aslett, an analyst with The 451 Group, told me, "Oracle's Linux strategy is about serving its existing customers," and, given that there are more Oracle customers using RHEL than Ubuntu--coupled with the fact, as Sean Michael Kerner points out, that Oracle has yet to certify its products to run on Ubuntu--RHEL (and its clone OEL) may be seen as the safer course of action.
Even so, it remains unclear why Oracle should continue to plow resources into OEL when the market is voting for RHEL (paid enterprise adoption) and Ubuntu (unpaid community and paid OEM adoption). Either go back to a full embrace of RHEL or try Ubuntu.
Oracle could turn that Ubuntu popularity into paid deployments, while simultaneously asserting a greater measure of control over its operating system story. I'm a big fan of Red Hat's business, but I'm surprised Larry Machiavelli (er, Ellison) hasn't sought to check the ever-growing power of Red Hat in enterprise infrastructure.
What do you think? Would Ubuntu be a good move for Oracle, or is Linux such an afterthought for Oracle that the status quo with Red Hat is the right course of action?
Follow me on Twitter @mjasay.
Matt Asay brings a decade of in-the-trenches open-source business and legal experience to The Open Road, with an emphasis on emerging open-source business strategies and opportunities. Matt is vice president of business development at Alfresco, a company that develops open-source software for content management. He is a member of the CNET Blog Network and is not an employee of CNET. Disclosure. You can follow Matt on Twitter @mjasay. 
Genuinely, if there was any play to be made, I would think controlling SUSE would bring the biggest advantage. It's a similar platform, competing in the same space as Redhat, but the culture would fit with Oracle much better than Ubuntu. The problem is it would rock the boat with Redhat, but I guess it's obvious Oracle isn't completely averse to doing that, look at OEL, I guess Redhat and Oracles relationship is Unbreakable.
Secondly, I think Ubuntu is rising because of the community and I think if Oracle were to get involved it would be like Sun buying MySQL, but worse. I can't possibly see the positives outweighing the negatives with a relationship between Oracle and Ubuntu. Different niche, different culture, different profit margins.
However, enterprise users of Ubuntu would use Oracle databases if they were using them already as Windows user previously.
The elephant in the closet is health care. The entire health system is due to start the transition to electronic records, and each hospital, clinic, central repository, and physician's office will need a database.
Health care is considering inexpensive open souce solutions, but if Oracle positions itself as a player, it will be a strategic plus in this multi-billion dollar market.
It can partner with Canonical to provide an affordable turn-key healthcare solution and get in on the ground floor.
Or it can miss the boat and watch PostGRE or even MUMPS databases dominate healthcare solutions.
There are currently two bills in Congress to support open-source in health-care -- certainly Oracle isn't blind to that. But if it buys an OS, it may be open-source, but not GPL, and the community will turn a blind eye to it.
Oracle must tread lightly. The best it can do is the Zimbra model of Yahoo (but even that is unlikely for Oracle). It could simply make its product available to Linux users of any flavor (as a binary or open-but-proprietary-licensed source), ala Adobe Real Player, or sell its product for a competitive price through Canonical's distribution channels, as Canonical already does with Landscape.
RHEL's market is largely US-based, but Ubuntu is worldwide and grows faster overseas than in the US. Providing Debian/Ubuntu/Kubuntu compatible packaging will have better penetrance worlwide for Oracle with Ubuntu.
Interesting and good point of view, the question is that if Oracle is willing to do it.
I would be glad to see Debian GNU/Linux and Ubuntu certified and officially supported by Oracle. When talking about Ubuntu, please don't forget Debian on which it is based on please, always remember that Debian stable branch is rock solid;-)
Terry@Debian;-)
Oracle wants to increase is visibility in the growing Linux world, and when entire governmental organizations convert to Linux (such as in Germany, the French police system, and other well-known large scale enterprises), it does not want to be abandoned for MySQL.
Hence this article, which perhaps was seeded by Oracle (no tech article exists in a vaccuum). Very amusing. Oracle has always been very difficult to implement compared to MySQL (or PostGRESQL) and it realizes it needs to become more accessible.
I get it. Linux drives evolution (if not sales) in the OS world, and now it will drive the evolution of Oracle. Very, very interesting.
*Gasp!* And not have omnipotent, iFisted-control over the kingdom?! Perish the thought!
We all know Cnet loves Apple, when I look at their RSS feeds there are more iPhones than any other cellphones, even with many Android phones coming.
BTW, just because Steve Jobs didn't put Linux on a Mac does it mean it's not good?
Err, something to ponder: Solaris and AIX are server products... Windows concentrates mainly on the desktop (yes they have server products too, but seriously - the word 'Windows' usually translates to 'desktop' in most minds, IMHO).
The upshot was the SuSE box couldn't unmount the NFS share which consequently meant it couldn't shutdown.
My point? Even with Linux, unless I've tested things in advance I like to stick with distros from the same people which is probably why NASA use Fedora as their desktop distro with RHEL on the servers. RHEL Workstation will work best with RHEL Server (advanced or otherwise), SLED will work best with SLES and if Ubunutu's going to get on the desktops of businesses they'll probably be more inclined to use it if there's an equivalent server version, so long as it does everything they require of it.
If Canonical wants the support business and the funding for the continued development of the desktop distro they need to make Ubuntu an end-to-end solution as much as is possible without overextending themselves. Looks like they're way ahead of all of us on this one.
Too bad Linux can't make as much ads as MS or Apple... or else it'd be much more popular.
FWIW I Do use ubuntu, but actually modified Ubuntu(realtime) DAWs and win64DAW and my storage servers are currently Solaris10 like another comment I think Ubuntu needs to do a little more work on their servers I got fed up with every new release breaking SMB
And yes, I would love to see Ubuntu do to the desktop what Firefox has done to the browser. But only if wishes were horses...
Er, no. Canonical is most definitely a for-profit company (not that they have actually made any profits yet).
Your link to justify this oft repeated statement is pathetic.
Its a 2 years old blog and uses scientific methods like which web site gets more hits or distrowatch ranking (same thing) or which distro has more blogs (rabid fanbois) devoted to it....
Utter and total garbage.
I want numbers, not guesses.
Go to Engadget and Gizmodo: more than 50% of the articles are about apple products which means that mac's share arent 5% but 50% of the computer business using the 'they talk about it more' logic.
Repeat something often enough and it becomes the truth.
Im not saying that Ubuntu isnt the top distro at this moment but I have yet to see anyone give decent Linux numbers period, never mind differentiating between distros.
Feelings, opinions, hunches and gut instincts mean nothing.
And they certainly do not justify 'clear cut choice' or 'dominating'.
And mentioning your company's experience might be fascinating, it does in no way indicate desktop Linux adoption even though you throw numbers out.
Take a look at the Fedora Project's statistics pages.
https://fedoraproject.org/wiki/Statistics
Fedora is making a best effort approach to establish a set of solid trendable benchmarks associated with Fedora project activities in a number of ways. Not just client downloads but also tracking the install base looking for package updates via aggregate MirrorManager log analysis.
If other distributions picked up the MirrorManager technology they'd be able to provide equivalent side-by-side statistics to compare with Fedora's. And since MirrorManager is an open source web services project, there's nothing stopping them from working on integrating such a service into their package management system.
-jef
You "saw" numbers? On someone else's screen? Is Canonical planning to publish those numbers and how they got them? Do you enjoy being Canonical's heresay paddler?
Either publish the numbers and the methodology used to obtain the numbers...or don't talk about them. Its not a difficult rule to follow.
-jef
In spite of the fact it is not so popular in the community area, solaris is still one of the main choices for those who deploy oracle. And considering it already has full compatibility with oracle software and oracle is using a lot of solaris-specific features - why can't it become the OS of choice for the db company?
Purchasing the TruCluster source code off Compaq was a stroke of genius on Oracle's part. It enabled them to include the functionality they needed directly into their own product and in doing so, make the likes of Oracle RAC clusters (Oracle Parallel Server back then) platform agnostic. So no one OS can claim any special privileges with respect to Oracle.
Nexenta capitalises on Debian code and Sun's Solaris kernel (with the highly regarded ZFS file system). The OS allows to benefit from the extensive Ubuntu repository.
I would like to point out that at least Oracle Database 10g XE has been certified on Debian GNU/Linux and Ubuntu long time ago.
Interesting and good point of view, the question is that if Oracle is willing to do it.
I would be glad to see Debian GNU/Linux and Ubuntu certified and officially supported by Oracle. When talking about Ubuntu, please don't forget Debian on which it is based on please, always remember that Debian stable branch is rock solid;-)
Terry@Debian;-)
Ubuntu is on a six-month release cycle which is tied directly to Gnome's release cycle. Oracle's customer base, again, is too conservative to run upgrades every six months because the whole distro is set on a timeline related to a development of one particular desktop environment. These customers may be more willing to run Ubuntu or similar distros at home, but not in production. They have too much at stake to care if they're at version 2.4.2 of Gnome or 2.6.
Ubuntu LTS doesn't have a lengthy support cycle (three years desktop, five years server) compared to other distros, particularly those like RHEL, CentOS, Oracle Unbreakable (get the name right) Linux (note they're all really RHEL), which appeal to enterprise users. Whatever bandwagon appeal Ubuntu has among desktop users, it hasn't crossed over to the server side of things. And it won't, at least where criticality is involved. And that happens to be where Oracle butters their bread.
No, Google's stratergy (Linux & otherwise) should be Ubuntu.
The Lean, mean & green, Chrome OS for the light tasks like email, docs etc,
And Ubuntu for other CPU intensive stuff like CD burning etc etc.
Ubuntu is the BEST LINUX Distribution and I been using Linux for a long time, used Slackware for a long time too, but in my opinion they fell behind and I lost interest in their distribution.
- by NeonDiet July 21, 2009 3:23 PM PDT
- There's one link in the chain that's been ignored and is an obstacle to Ubuntu Server adoption in the enterprise, and that's platform support and certification from the hardware vendors. For example if I want to install HP's Proliant Support Pack on Linux, I can find a kit for RHEL 4 and 5, SuSE 9, 10 and 11, and even Debian (lenny) but not Ubuntu. There is just no way that I would deploy an HP server without for example the benefits of HP's patched drivers, the ability to monitor system health and manipulate onboard RAID controllers from within Linux using PSP supplied tools. It doesn't matter how much blessing Oracle heap on Ubuntu.
- Like this Reply to this comment
-
(35 Comments)And what version of Ubuntu would Oracle and the vendors choose to support? The LTS releases, ignoring those in between? 8.04 is getting old very quickly now, but 8.10 and 9.04 have not exactly been shining beacons of stability. Ubuntu's insistence on fixed 6 monthly releases has exposed their user base to bleeding edge **** ups that have tarnished Ubuntu's image of late. For example their decision to push UXA in 9.04 before the Intel Graphics drivers were ready for it caused serious regressions in performance and stability for just about anyone with an Intel Graphics chip (me included). I'm not sure I would trust them with a mission critical server after that, and so far the lack of hardware vendor support indicates they have yet to be convinced too.