Where Red Hat (And Its Partners) Profit Most
Ever wonder why Red Hat spends so much time focused on the JBoss middleware market and so little time trying to make Linux a desktop standard? The answer involves some simple but startling open source math. Check out this little piece of Red Hat financial info, uncovered by The VAR Guy.
Here’s the simple formula: For every dollar a customer spends on JBoss (the product), that same customer typically spends another $10 to $12 on related JBoss consulting and integration services, notes Mark Enzweiler, VP, global channel sales at Red Hat.
“JBoss middleware is like a central nervous system,” asserts Enzweiler. It’s so mission critical that customers are willing to pay Red Hat or the company’s integration partners roughly $11.12 in consulting fees for every $1.00 they spend on JBoss itself, estimated Enzweiler.
In stark contrast, Enzweiler hasn’t been able to make a strong business case for signing Linux desktop OEM agreements.
Sure, Canonical’s Ubuntu Linux and Novell SuSE Linux (to some extent) earn headlines each time they win a new business deal. But during preliminary OEM negotiations, Enzweiler determined that PC companies “basically wanted it [desktop Linux] for free.” At such low prices, Enzweiler worries that Linux distributors may wind up compromising their support models to gain market share on the desktop.
Still, The VAR Guy wonders: Should Red Hat find a bit more time for desktop Linux? Remember, Microsoft used its desktop position in the 1990s as a springboard onto the server. Canonical hopes to do the same with Ubuntu Linux.
But so far, Red Hat doesn’t appear concerned. Indeed, Enzweiler continues to evangelize server Linux — and JBoss middleware — to Red Hat’s integration partners. So far, simple math shows that Enzweiler’s strategy is a good one.
I think its a serious mistake by RedHat and Novell to let the desktop slide. I find myself using Ubuntu more and more in my server installations. The fact that i use Ubuntu on the desktop does indeed matter when i choose OS.
Server vendors trying to lock people into particular dists wont succeed in the Linux world so i dont think anyone trying that will fly for very long. Its more about familiarity and what people think works best. A very good desktop OS is mighty good PR for any server products.
Daniel: You could have a point. Short-term Red Hat will do very well with its server and middleware strategy. But the company will need to guard against desktop upstarts (particularly Ubuntu) sneaking onto the server.
By the way: The VAR Guy didn’t says Novell is letting the desktop slide. On the contrary, the company has won a desktop OEM agreements in recent months.
What pushed MS into the server room, it seems to me, were services that related to and completed the desktop experience — most notably, active directory (NT domains, originally). For Ubuntu to leverage its desktop popularity to get it into the server room, it needs to integrate client and server in some special and useful way. I don’t know if having an LDAP-in-a-can type solution like AD is still relevant these days, or if it needs to involve desktop virtualization or some other red-hot buzzword technology, but they need something that enhances or eases the management of Ubuntu clients.
alan: have you looked at Canonical’s Landscape tool for remote desktop management?
I don’t think Red Hat has anything to worry about from Ubuntu/Canonical. Enterprise users are content with Windows on both desktops and servers and are prepared to pay for middleware support directly from Red Hat when they do run Linux servers. Canonical simply doesn’t have the gravitas in that arena that Red Hat does — not yet, and probably not in the near future. Just look at the respective software under their control: jboss versus launchpad or bzr? Give me a freaking break. Red Hat isn’t competing against Ubuntu, they’re competing against Oracle, Sun, and other vendors who have a lot more to offer enterprise users than a live CD-based desktop distro that’s oriented for non-enterprise use (Ubuntu is too bleeding edge with too many security issues for most shops to take it seriously).
@TVG: I hadn’t, but it looks intriguing. Is it enough to gel the concept of an “Ubuntu network”, the way Domains did for Windows? And for that price (consider how many folks have sold Linux to the PHB on the idea that it’s “free” or “cheap”)?
Not trying to be a critic, but it seems that RH has carved out a niche as a replacement for Unix and a platform for Java services, and Novell is using Suse to succeed Netware. Neither approach really relies on a linux desktop solution; but as the desktop is Ubuntu’s strength they are going to have to make client integration the “killer app” for Ubuntu server.
And I gotta second lucky13; even though I use Ubuntu on my desktop, I have stuck with Debian stable for servers at work rather than make the move to Ubuntu LTS. When HH came out with Firefox 3 beta as the default “firefox” package, I lost serious faith in the judgement of the distro maintainers to give me stable software.
But I say all that as an armchair critic; I suspect Shuttleworth knows what he’s about. I am interested to see what a fully-formed “ubuntu server edition” shapes up to be.
Alan: Sounds like you have some healthy perspectives. The VAR Guy agrees with you on a number of fronts. But a few key points … Red Hat has more than a niche. They’re on pace to become a $1 billion company in a year or two. Novell SuSE Linux is far more than a replacement for NetWare.
Ubuntu faces an uphill battle on the server. But Mark Shuttleworth certainly is sharp, The VAR Guy agrees.
While the Microsoft-OEM cartel that keeps alternative OSes from breaking into the market using tactics such as price fixing coupled with advertising and other rebates, is allowed to operate unfettered, the desktop OS market will always be a loss making prospect for anyone other than Microsoft.
I think Ubuntu will do well with SMB’s on the server side but not do anything in the enterprise market. Red Hat has that market.
People forget how MS got a foot hold with their server.
1. They almost gave it away. As Banyan required server keys and Novell required license key files. Where Windows NT just required a license key. So you could roll out 10 Windows servers and pay for one license. That is not “legal” but MS didn’t care cause they got market share that way.
2. They marketed it against Novell and Banyan as more easy to use with Windows desktops. (No client)
3. Free patches. This at the time was a big deal. Banyan and Novell used to charge for patching etc.
4. Windows domains (At the time) were more easy to set up and use then LDAP (NDS and Banyan Vines at the time) So even though NT domains were not as secure and robust as NDS, they were easy to set up. You didn’t need network admins to support it. With Novell you needed a certified Novell Admin. With Windows you didn’t
Now everyone is hooked on Windows. People need Windows because its what people use at work etc.
Right now Redhat needs to keep focus. Don’t worry about Ubuntu. In the future maybe they could partner up to make an infastructure of Ubuntu on the desktop and Red Hat on the backend.
Interesting facts.I have bookmarked this site. stephanazs
I just dumped my redhat stock and this was part of the reason.
I’ve had the stock from close to the beginning of the company.
Redhat no longer has a vision, I didn’t hold stock for the short term I held it for the long term RHT does not appear to have a strong steady long term view anymore meaning NO VISION.
Ubuntu? Don’t know, don’t care I think it’s time for a new generation of young turks to turn things upside down again and rediscover Free Software and Intellectual property.
Linux has become too bloated and too encumbered and as an “older” technologist I’m willing to say, Linux has become too OLD, creaky and yes – broken.
PaulG: The VAR Guy doesn’t get into buy, sell or hold discussions. But the bigger discussion here is whether open source companies can generate long-term growth and returns for investors.
PaulG Says:
“I just dumped my redhat stock and this was part of the reason.”
Thanks for the stock, Paul. There’s a winner and a loser on each end of every transaction. I’ll add your stock to the rest of my holdings on Red Hat.
“Redhat no longer has a vision, I didn’t hold stock for the short term I held it for the long term RHT does not appear to have a strong steady long term view anymore meaning NO VISION.”
Silly rabbit. Redhat is skimming the cream. As it should. They and all other FOSS companies I can think of, but especially FOSS pure plays, simply don’t have the money nor the time to waste to reinvent a desktop that will displace the current monopoly. They wisely left that to other disros, other companies, and the community at large. Once someone has a sustainable solution _that_ _is_ _profitable_, then thanks to the GPL, they can step in and release their own version. In the meantime, they are concentrating on skimming off the very substantial Unix installed base that was out there the last few years and is still out there thanks to long term contracts. There’s no way they should concede this market segment to pursue a money loser like you suggest. SELinux, RealTime Linux, LinuxDevices, JBoss, there is a lot on their plate. Luckily they are being smart about it instead of being gluttons and trying to devote the money (that they don’t have enough of for this segment) and the resources to compete with the terminator when it comes to the desktop.
“Ubuntu? Don’t know, don’t care”
Yep, Ubuntu is a flash in the pan, and RedHat has no worries over Canonical. Ubuntu is a popular Debian desktop, and allows for an easier and newer install for servers. But it doesn’t exist in enterprise server rooms, and experienced Linux admins and CIOs will go with pure Debian for servers if they decide to go with a Debian-based distro at all. Thanks to server “policies” and software/update “policies” along with requiring a neck to choke, RedHat or Novell are the two options for CIOs, and CentOS or Debian are for the more flexible companies that allow some degree of experimentation. The number of companies that allow some degree of experimentation will grow over the next few years as the economy slows and belts tighten.
“I think it’s time for a new generation of young turks to turn things upside down again and rediscover Free Software”
There are plenty of distros out there for them to do this. Slackware is still alive and kicking. LFS and DFS are still available. In the meantime, colleges and Universities are either moving from or have already moved from HP/UX, AIX, VAX, Solaris, and are all teaching _on_ GNU/Linux and are all teaching GNU/Linux. The only exceptions are the die-hard professors pushing retirement age who lived and breathed Unix in their early years and who are living on vapors right now.
“and Intellectual property.”
Yeah. And this is where I call bullsh*t. Your post had a whiff of troll to it, but this is where you pulled your own pants down.
“Linux has become too bloated and too encumbered and as an “older” technologist I’m willing to say, Linux has become too OLD, creaky and yes – broken.”
Troll. Because Linux the kernel and GNU the apps are so flexible that it runs on a far wider collection of architectures than Windows or Unix does. Don’t know if it has surpassed NetBSD yet in terms of number of architectures it runs on, but if not, it will very soon.
As for its flexibility and nimbleness, and its modularity, GNU/Linux is running Google, the NYSE and other markets, it’s running on mobile phones, it’s running power grids, telecommunication and IP backbones, medical equipment, studio audio equipment, Hollywood, all or nearly all special effects studios’ render farms, the Internet, robotics, LCD televisions, DVRs, transit systems, freight systems, shipping systems, nearly any other embedded system you can think of…
Yeah, PaulG, that some bloated, creaky, encumbered, broken, piece of cr*p, that Linux. And I’m happy to report that the broken piece of cr*p as you call it has been happily running on my desktops approximately the last ten years.
Wow, Seven. A little ruffled there by the reference to “intellectual property”? “Property” got a bad name for you? See your shrink before you have a heart attack! The term “Troll” was coined and promoted by Intel, to disparage small legal shops who would work on behalf of small inventors–the young turks–to try to crack Intel’s IP fortress.
Intellectual property is at the core of innovation: good ideas are assets. Whether or not the inventors/owners ultimately decide to share their ideas (old school open source) and/or allow free distribution of their code (as in free beer) makes no difference to the value of the property.
In general free and open source software is software built on a core that has been commoditised, and that has little value as property. Firefox is the old Netscape code, Linux is essentially commodity Unix.
Successful companies like Redhat don’t succeed because they distribute the commodities–they succeed because they build on these commodities highly differentiated offers combining middleware, certified applications, and an ecosystem of experience and trustworthy vars.
OS per se are just not that important anymore. Almost all the innovation and the willingness of customers to pay–has shifted up the stack and out into the ecosystem, and ultimately into networks of web services.
I think you are missing something very important about Unix/Linux GNU/Linux–the reason it is running in all those places is the same reason Apache is–both elements of the stack are fully commodities, and very little value is added in either. This is also the reason you have been able to run it happily for so many years. At base, it has not had to evolve, because the evolution happened around it, not within it.
Case in point: Google’s use of Linux is not an endorsement, it is an appropriation. Google could care less what it runs as an OS. What it is focused on is its network of advertisers. Oh yes, and its massive and growing collection of software and services patents. Up the stack and out into the ecosystem.