Red Hat Dismisses Consumer Desktop Linux (Again)
Sometimes, you have to respect someone for sticking to an opinion and a vision. Other times, you have to wonder if a consistent vision becomes a fatal flaw. I’m still undecided about whether Red Hat CEO Jim Whitehurst has the correct — or flawed — long term vision: In a New York Times article, Whitehurst (pictured, right) once again dismisses speculation that Red Hat will pursue the consumer Linux desktop and smart phone markets. Here’s why.
“We have an extraordinary brand, and the last thing I want is for someone to take a computer home and find something is not working,” Whitehurst told The Times.
Short term, it’s difficult to argue with Whitehurst’s logic. Many consumers wind up disappointed after they purchase Netbooks without realizing Linux is running behind the curtains.
Long term, I’ve got to think the consumer Linux story is going to have a happy ending. And eventually, as Canonical grows up, Ubuntu will make the jump from desktops to mobile Internet devices and servers — putting the squeeze on Red Hat in selected areas, such as small business servers and departmental servers.
Red Hat is putting all of its eggs in the server basket — while branching out into JBoss middleware. Is Whitehurst brilliant — or too stubborn to open his eyes and his mind to the desktop opportunity?
WorksWithU is updated multiple times per week. Don’t miss a single post. Sign up for our RSS and Twitter feeds (available now) and newsletter (coming in 2009).
WorksWithU on Identi.ca
WorksWithU on Twitter
WorksWithU's RSS Feed
The NYT’s story is a non-story. Red Hat is not in the consumer space and is not aiming for that space, but is profitable in servers. So? The bigger story is what, if anything, Linux supporters, developers and hardware makers are doing about the “What’s this? Where’s my Microsoft Word?” factor. It is a profound PR and marketing failure that consumers taking home Linux netbooks are surprised and wish to return the product.
I used to run Red Hat Linux back in the day but since I have moved to openSUSE, Ubuntu and my fav, Kubuntu 9.04. Ubuntu is going into small mobile internet and into the server market and it will leave Red Hat in the dust soon…
I’m not a “techie” I’ve been using Ubuntu desktop since the first LTS. Before that it was Debian, and before that Redhat (before Fedora). Once I associated Linux with Redhat, mainly because I had been using it since version 4.something. Now I associate Linux with Ubuntu. I know Redhat probably makes the best server and server applications, but, when I want a server now I reach for my Ubuntu disk because I’m familiar with the “name” and the forums. A lot of people that are just starting out in servers probably do what I do … it’s a name they are familiar with. That will someday hurt Redhat.
/*“We have an extraordinary brand, and the last thing I want is for someone to take a computer home and find something is not working,” Whitehurst told The Times.*/
I guess Apple is doing something wrong….
redhat is human computing. redhat is like humans. humans learn nothing from history. 50,000 years back they fought and killed each other and do same today, just new techniques. redhat is similar, following foot steps on novell and solaris, only servers. any body remembers novell, died and solaris is dying. these days only enterprise apps care about solaris, internet apps is linux.
Thats right. Thats all Ubumtu is, a name. Someday they might do something to ooh and ahh the real Linux people. But until then I’ll stick my Mandriva.
And yet… Red Hat is continually making development investments into upstream desktop project components as part of GNOME that directly impact the viability of an open platform in the consumer space. Something Canonical is not doing. Care to count how many Red Hat employees are listed as maintainers on GNOME components? Hell, I think Litl as a company has more employee activity in upstream GNOME development than Canonical does in GNOME…and litl is super super tiny. If the consumer space matters to you, you should keep your ear to the ground concerning litl. I guess we’ll see what litl brings to the table as soon as they come out of stealth mode.
http://blogs.gnome.org/lucasr/2008/04/14/joining-litl/
http://log.ometer.com/2009-03.html#22
You want to think long term? Long term is all about directly supporting the development roadmaps of upstream projects like GNOME, like X.org, like the kernel. And that is exactly what Red Hat does. You take a good hard look at the make up of the development roadmap and the commit logs of GNOME or look at the accumulated activity around all the software and specifications associated with freedesktop.org and you’ll see very quickly exactly what sort of long term commitment Red Hat makes to the linux desktop. The commitment is there, the work on things like HAL/DeviceKit or NetworkManager have nothing really to do with their core server market and everything to do with the long term viability of consumer linux.
The development work Red Hat employees do for desktop oriented technologies enables everyone else to chase the consumer markets..because its done as part of upstream project development. The question isn’t whether or not Red Hat supports consumer linux. They do, they are right in there participating in the development of what is coming next. They are doing far more than their fair share of the work considering the desktop isn’t a critical product market for them. No, the real question is, can they sustain a product offering in the marketplace. Can anyone?
More likely than not consumer linux is going to only be sustainable long term when the hardware manufacturers get deeply involved in the development of a linux operating system for themselves. In much the same way that Apple is really a hardware company..that produces its own operating system as part of the overall device design..so too will harware vendors who will eventually be able to sustain a long term linux development effort. Long term, the investment Intel is making into Moblin is going to have the highest impact on the viability of consumer oriented linux offerings, more so than the work Canonical is doing. I think the evolution of Moblin as an distribution agnostic project is quite telling. I think as consumer hardware vendors gain more confidence with the linux ecosystem they are going to prefer more direct input into development roadmaps of individual components instead of paying for an integration services vendor, like Canonical. People who purchase consumer devices certainly aren’t going to be paying for software support contracts as an addon cost.
Canonical’s development efforts into the desktop to date aren’t very interesting as they’ve done an extremely poor job of working with upstream projects to integrate those changes into the upstream development roadmaps. They are probably doing a far better job on the server front in that regard. The ecryptfs work Kirkland is doing made it up into the appropriate kernel subsystem tree quite quickly. Contrast that to the amount of downstream patches Canonical created for presence featuring in Intripid and now for notification featuring in Jaunty. Canonical may feel justified doing a lot of downstream development work on the desktop that has no path towards integration in the upstream projects and Ubuntu users may like to see those features…but in the long term its going to be a problem as upstream projects implement their own competing solutions, obsoleting Canonical’s work. That sort of behaviour is not long term thinking, its short term.
-jef
Red Hat tried to sell boxed linux at retail stores back in the olden days of Linux. That failed financially, even when broadband was rare. They want a powerhouse consumer channel before tiptoeing into that market again. They are keeping that option open with Fedora and still getting value from the expense as a test platform for RHEL. I think they’ll come back when someone makes them an offer.
An offer like: “We’re going to sell 100,000 of these machines with Fedora desktop on them. If you’ll provide the spiffy-shine to make it a great consumer product, we’ll make sure you have test machines to refine it. If that market works, we’ll probably do 500,000 in the next iteration. How about it?”
Who says that every Linux vendor has to have a desktop product? Does every Linux vendor have to cover every aspect of computing with their product? How about embedded, mobile, realtime, or cluster OS configurations? Do Novell, RedHat, Canonical, et al have to cover all those bases, or are they allowed to specialize?
Let RedHat do what it does best.
You take the good along with the bad (or not-so-good). I think Whitehurst is being shortsighted. By doing the server only, Red Hat is missing a learning opportunity. Sure, Microsoft is already in the desktop space. But Red Hat is not alone in the quest for the desktop, there’s also the Linux community behind it. I think Whitehurst keeps forgetting about the community. The community is more than up to the challenge, the fight.
The only problem with what Whitehurst said previously is that he seemed to diss desktop Linux in general. If I remember he said something to the effect that desktop Linux isnt going anyplace. I think they are smarting from the Ubuntu effect.
BR Jones, Ubuntu isn’t going to “leave Red Hat in the dust” anytime soon given Ubuntu’s buggy releases and overall flimsy and rushed development process.
If they want to compete with RH head-on they need to create a much more refined and stable distro….BUT, why bother since that is what Debian already provides!
I don’t know why so many people think Ubuntu is a one-size-fits-all solution to a variety of problems.
“I guess Apple is doing something wrong….” One of the things going for Apple, is that they only have to support their own hardware. What Red Hat wants to avoid is all the problems coming with random hardware… so much of it doesn’t work well with Linux, or acquire a fairly hefty dose of time and knowledge to make it work (BTDT – I’ve been doing linux since ‘95).
The server market is much more controlled – you don’t plug in something cool you bought at walmart. The next logical step would be the corporate desktop, where you also have much more control of the hardware. Consumer desktops, if ever, is a market Red Hat can leave alone for a while longer.
Ken Harbit
You hit it right on the head. What Red Hat is doing is trying to play down Canonical as much as possible. Its all they are able to do. If Ubuntu finds its way into the mainstream and people become familiar with it then the server will follow.
I’m an example of it. I used to be a 100% Red Hat/CentOS server guy. But after using Ubuntu on the desktop for 2 years I have decided that my next large project will use Debian or Ubuntu. Just as with the desktop when I used Fedora I started finding too many of my answers in the Ubuntu community so I switched. Now with servers I’m finding answers in the Ubuntu/Debian community but more importantly I’m finding the packages I need. I end up spending alot of time trying to find the packages I need for RHEL when a “sudo apt-get” gets the job done while I’m developing on Ubuntu and now testing on Debian.
Red Hat is not being short sighted. They are actually looking down the road at the results of an opportunity they missed. All they can hope to do is help Canonical fail. Surely you can’t believe Whitehurst’s suggestion that you can’t get everything working that a consumer needs on a Linux desktop can you?
I can understand how he might feel. There was a local electronics store that carried computers running Linspire Linux. Every time I looked at one, it was connected to a monochrome monitor, and misconfigured to the point of absolute uselessness. The reason for this, I found out after complaining to them (and Linspire as well) many times about it, was because their sales people had no clue how to work with it, and decided to make it look “broken” to discourage potential customers from asking questions about it. I finally sent a letter to their home office saying that if they didn’t want to sell it, they ought to stop carrying it, and stop making themselves look like idiots. Eventually, they did.
While I agree with Ken and Ron and others here that Red Hat’s approach will hurt someday, I can understand that Red Hat would stay out of the current desktop mess.
It’s comfortable and stylish to pat ourselves on the back about how great desktop Linux is, but there really isn’t much objective evidence that desktop Linux will soon be a big success.
For every technically-capable computer user who is happy with Ubuntu, there’s a common consumer who will never try Ubuntu again. This equation will change someday, but not tomorrow and maybe not soon.
It’s no use going into the reasons cause nobody ever got anything but roasted for criticizing desktop Linux on a Linux site. Surely a thoughtful person can see the reasons and can understand that Red Hat wants no part of it at this point.
Until money start changing hands linux on desktop will continue to be a failure. And Red Hat understands it, so does Shuttleworth
It is possible that Jim Whitehurst’s fixed vision may turn out to be “a fatal flaw” IF Microsoft finds an unbreakable way to control servers and middleware through some new feature/capability link to Windows Desktop, and Microsoft Exchange.
This also assumes that the mobile market will continue to hold Microsoft influence at bay with Google Android, Symbian, Apple IPhone technology and other superior products.
Maybe they can even do same through Lawsuits and Patent claims. They certainly will try anything.
W. Anderson
I see a big difference in the vision between Mark and Jim. I’m with Mark on this.
Redhat could be downplaying Canonical or the desktop or both. But if Ubuntu keeps gaining momentum, by the time Red Hat decides to join the desktop market it may be too late.
The success of RH’s tactic will remain to be seen. In the business sense, they are sticking to what they know best, enterprise servers and middleware. Their knowledgebase is concentrated on this area meaning they can be a bit more efficient rather than dealing with finicky desktop users – and believe me, they can really ruin a techie’s day.
With such a conservative approach, they are still contributing TONS of stuff to the desktop environment, probably waiting until they can sell a total drop-in Linux environment. The pieces are there, and I have a feeling they are waiting for the glue.
RedHat aims to the unix replacement market… its clear strategy which make that small company successful (in terms of global market)…
Product line covering consumer desktop operation system is Very Expensive in development and especially service…
Red Hat lacks such resources needed to make it profitable right now however it doesn’t mean they have no plan for the future…
“We have an extraordinary brand, and the last thing I want is for someone to take a computer home and find something is not working,”
There’s the open source spirit. Everybody pitching in to make it work.
Frankly I think that’s a lame excuse. Never was it written than an OS had to work with absolutely every single piece of kit out there know to man and beast and earthworm. In fact I can’t name a single OS which does work flawlessly with every single device. Even in the consumer space.
The approach Linux distribution developers should be taking to desktop Linux is exactly the approach Canonical has taken with Ubuntu. Get involved with a few top tier hardware vendors and guess what? People start to make their stuff work with your stuff! It’s like magic!
My position on this keeps changing. I think Red Hat is in danger of missing the boat on desktop Linux. But many readers have pointed out that Red Hat contributes quite a lot to the desktop FOSS effort without necessarily targeting consumer Linux pre-installs.
Red Hat need to make contributions to keep their army of Fedora testers happy. Without the Fedora community Red Hat would have to do an awful lot more in house testing which would obviously hurt. And lets face it, there’s little point in being an open source company if you’re doing it all your self. So it’s give and take just the way FOSS should be. Everybody goes home happy.
If Red Hat don’t want to push out a desktop specific OS that’s fine. They have a strong server OS business with the services to back it up. What I don’t like are the lame excuses.
Where Red Hat will lose out is in visualised desktops designed to work with thin clients and netbooks (think mobile thin clients booting over a VPN). IBM are already doing something similar with Ubuntu.
@aikiwolfie
Yes lets talk about virtualized desktops. and that sweetheart deal between IBM, Canonical and Virtual Bridges.
The announced product from IBM in December was really a veiled announcement for the Virtual Bridges VERDE VDI technology.
http://www-03.ibm.com/press/us/en/pressrelease/26230.wss
IBM injected a cut down version of lotus suite (occs) into that offering at no-cost as a value add. The bulk of the per-user pricetag, $49, went to virtual bridges. $10 went to Canonical for support. That’s a pretty sweet deal for Canonical and for Virtual Bridges. But that was december. Is that product offer still on the table at the price from virtual bridges?
I look at virtual bridges website today and I see a $125 per user with a minimum of 10 users:
http://vbridges.com/wp/?page_id=108
okay actually its a little less than that for the first 10 users.. just under $1k
http://vbridges.com/wp/?page_id=96
I guess you have to contact virtual bridges sales and make a direct inquiry with regard to that special $59 per user deal from December. If that Ubuntu based VDL solution was really getting strong support and interest, you’d think they’d have it placed prominently on their product page. As it stands it sounds like that price point was a limited time promotional offer from virtual bridges. Maybe someone here at workswithu can contact virtual bridges and ask them why the $59 per user price point from the december announcement isn’t listed in their online store. Is Canonical still getting a cut of the $125 price? Or are consumers going to have to contract with Canonical separately for support?
You’ll also note that Virtual Bridges extended VERDE to work with RHEL as well:
http://vbridges.com/wp/?p=86
Maybe that’s because Red Hat is still recognized as the established leader in virtualization:
http://press.redhat.com/2009/03/24/red-hat-enterprise-linux-collects-industry-awards-for-leadership-in-virtualization-green-it-and-overall-platform-value/
So what about IBM? IBM has historically tried to support multiple linux vendors at that same time. They do not want to make the mistake of creating another Microsoft by just supporting one linux vendor. And it is a smart move on their part. But don’t mistake IBM’s support of Canonical as a more than it is…it is market place competition maintenance. IBM keeps the door open for Canonical to find its feet, just like they keep the door open for Novell to find its feet…but its sure not an exclusive love fest. IBM isn’t going to get stuck in an exclusive deal with anyone. They are not going to pick a defacto winner. They are going to continue to throw vendors into the ring and let them fight. Look at this page:
http://www-01.ibm.com/software/lotus/openclient/
links for Red Hat, Ubuntu and Novell all listed as occs partners. Here’s a press clipping that summaries IBM multi-vendor support for occs:
http://biz.yahoo.com/iw/080805/0422227.html
Maybe you want to compare and contrast these two pages:
http://www.redhat.com/partners/partnerspotlight/ibm/lotus/
http://www.ubuntu.com/partners/occs
Anything significantly different between the two?
-jef