Sometimes I don’t see eye to eye with the Ubuntu user community. For instance, I think Ubuntu Server Edition should be Canonical’s top priority for 2009. But only 8 percent of WorksWithU’s readers agree with me, according to a recent online poll we ran. Here’s why I’m right and 92 percent of WorksWithU’s readers are wrong.
Consider the following scenario:
What if Microsoft had never moved beyond the desktop in the 1990s? What if the software giant never built Windows NT Server, SQL Server, Exchange Server and all of the other Microsoft server applications?
Fact is: Starting in the mid-1990s, Microsoft’s server efforts made the company far more profitable — and far more strategic to businesses and enterprises, and the Windows Server applications suite tightened Microsoft’s grip on business customers.
Changing Software Rules
Fast forward to the present day and desktops are becoming LESS strategic. Netbooks, PCs and mobile devices are commodity products that plug into the Internet and virtualized data centers. We spend all of our time living in Web browsers rather than desktop operating systems.
Yes, there’s still money in desktop and mobile software — especially as smart phones and netbooks proliferate. But if Canonical — and the Ubuntu community — really want to succeed, the road to long-term success and profits begins on the server.
Canonical has been working overtime to attract more ISVs (independent software vendors) and hardware makers to Ubuntu Server Edition. The company is making progress. But stay tuned: More progress is coming soon…
Reader Poll
… in the meantime, I continue to disagree with our latest WorksWithU reader poll. We asked: What should Canonical’s top priority be for 2009? Readers responded as follows:
- 41 percent: Polishing Jaunty Jackalope (Ubuntu 9.04) for April launch
- 27 percent: More consumer and business PC preloads
- 16 percent: More notebook and netbook preloads
- 8 percent: More server hardware and ISV relationships
- 8 percent: Recruiting and training VARs, integrators and consultants
Of course, Ubuntu polls will tend to favor desktop outcomes and desktop priorities — since the vast majority of Ubuntu users focus on the desktop.
But I still maintain: Canonical’s top priority should be server and ISV relationships, followed closely by VAR and channel partner relationships. Those are the key audiences that will make Canonical a viable business for the long haul. Again, stay tuned.
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).
Read More About This Topic
Share This Post
Interact: Add a Comment | Trackback Link | Permalink
Subscribe: RSS Feed
Don't miss Charlene O'Hanlon's weekly columns...
Netbooks are a huge break for desktop Linux, Ubuntu or otherwise. But desktops still need work to make them as seamless as Windows is perceived to be. If Canonical should focus on the server market, who would you say should focus on the desktop?
Scott: I think Canonical should continue desktop efforts but make server the number one priority.
This is *exactly* what Red Hat said a few years ago, and why Red Hat Linux shrinkwrap boxes immediately disappeared from stores around the world, to be replaced with online-only Fedora which struggled for years and still is nowhere near as popular as Ubuntu on the desktop.
Ubuntu is finally doing desktops right, and any attention drawn away from that is going to forever delay “Linux on the desktop”. Red Hat and non-Linux vendors like Sun are already doing a good job on server software and support, and Ubuntu just doesn’t have the manpower right now to retool to server software engineering.
In fact Ubuntu has yet to release a single version which is stable enough from an enterprise server point of view, and it seems to be the case that this cannot be achieved with their release engineering strategy at all. Debian and Red Hat have produced incredibly robust releases through a much more conservative release engineering strategy. Red Hat supports its own releases, but maybe Canonical’s best bet, in all seriousness, is to offer commercial support for Debian server installations. That would be a lot more effective than trying to hype up an OS which is already buggy on the desktop as somehow adequate for a server.
Well… I voted in selfish reasons, as I think did a large majority.
I was thinking about the ‘wrong market’ thing before – but on a different aspect, and that is making it programmer-friendly. Microsoft really “got it right” with MSDN – and Ubuntu is nearly looking in the opposite direction, having put documentation for the apps it made into various wiki pages / source packages / bazaar code branches.
Gnome at least is doing a better job of being developer-friendly (http://library.gnome.org/devel/)
“We asked: What should Canonical’s top priority be for 2009?”
Well, your argument appears somewhat circular.
You poll mostly Ubuntu DESKTOP users (which are the majority of Ubuntu users) about what they want, and seem surprised they don’t focus more on the Server? Seems kind of a foregone conclusion that Ubuntu users being mostly deskop-users, they would have desktop-related improvements top of their list, doesn’t it? What the heck did you expect? If you ask a desktop user what they want, why would you expect them to respond differently?
Now, had you polled Ubuntu Server operators what the top priority should be made, you might have had more Server-oriented responses. And they would probably have supported your own position.
Your conclusions are based on very faulty logic. Your logic.
regards,
-dc
Deadcats: I’ve got a hidden agenda. We’re going to attract Ubuntu server administrators to WorksWithU by the truckload over the next year or two. I applaud Canonical’s momentum on netbooks and Dell’s desktops. But to me, this market needs much, much better coverage of Canonical’s server moves. And we’re going to deliver that server coverage while continuing our desktop focus.
So, you are right: I used faulty logic. But I think the end result will be more people talking — and thinking — about Ubuntu server.
Regardless, we welcome your constructive criticism. Keep it coming and thanks for reading WorksWithU.
imho: i think netbook should be the priority.
1. this is where Microsoft bleeds, if you want to hit MS where it hurts!
2. GNU/Linux doesn’t have many killer applications and minor printing issues that make it hard to be adopted in corporate level, particularly for clerical task, while fit for network admin. Netbook doesn’t need those, combined with its hardware that’s more suitable to linux than MS, this is the area where ubuntu, along with its popularity, will shine.
3. i see new trend that netbook will become like cellphone. Blackberry is nice, but it is not fit for browsing and many handy use that netbook can do. There will be many netbooks in a few years to come, especially in schools and campuses.
budhi: I respect your opinion. Netbooks are hot and they are impacting Microsoft’s profit margins. But ironically the netbook boom has potentially hurt Canonical’s own Mobile Internet Device (MID) strategy, which was supposed to involve touchpad-like devices that look like smart phones but depend mostly on WiFi.
Regardless, I’m sure Canonical doesn’t mind that its netbook success has hurt the MID initiative so far. I’m still screaming “server” as a priority, but I like your three points for netbook. They’re well-stated.
I’d stick with both. That’s what Microsoft always did. And they got nice profits in the small/medium business world.
Joe,
Its really none of the above. Ubuntu Server is already a decent product for the layer it occupies. So there is but a few areas that I would expect to see an improvement in to make it great.
But take it up a layer and what does Windows have that Ubuntu does not but that medium to large business’s demand? A) Rock solid enterprise authentication B) Top notch enterprise directory services. C) Easy to use mgt console so that the task doesn’t take a rocket scientist. Another words Active Directory. (And yes I could cobble together Kerberos and LDAP eDirectory but that is not integrated.)
Take a look at any survey of the last 5 years. Linux does quite well up to about a 500 seat installation as the primary backend engine. Linux has the startup and low end of the SMB market locked up. Beyond that Linux starts to fade and Windows tends to take over.
Till a FOSS product comes to the fore that gives AD a run for the money Linux will participate in but not lord over the enterprise. Now Ubuntu has that problem as well. So what to do? Well Samba4 is going to be released soon. The rumble is that Samba4 can be a plug replacement for much of the AD function for most windows clients. So Canonical might find it worthwhile to fill in any holes in the Samba delivery and develop management wrapper code to run it all. That way Linux can sit at the top of the enterprise food chain than down at the middle layer. Doing so gives Canonical the opportunity to move into the larger seat count installations.
JohnMc: We’ll be sure to keep Samba4 on our radar. Thanks for weighing in.
I agree with your assertion, but I would say there is a ton of effort going into Ubuntu server. Canonical and the Ubuntu Server community are really building an incredible amount of new infrastructure functionality into the server. The ability to scale from the ‘local’ datacenter to the cloud, more tightly integrated server services, tighter integration with AD via likewise, and much more. The virtualization is already superiour. Jaunty is an exciting release on the server side.
Dan: We’re on the same page. Canonical deserves huge credit for moving into highly competitive markets. But I also think the company is stretched thin between multiple initiatives.
All that said, I dropped a few clues in the blog entry saying “stay tuned.” This blog is part one of a two parter. And part 2 is going to be huge. Stay tuned
“The ability to scale from the ‘local’ datacenter to the cloud, more tightly integrated server services, tighter integration with AD via likewise, and much more.” — dan
Seems dan has made my case. Linux can participate in a Windows environment. Likewise is but one tool in that arsenal. But why shouldn’t Linux supplant a Windows environment?
Hi there,
from my viewpoint, partly you are right and at same time you are not.
At the point when Microsoft started concentrating on server market, they were already defacto standard on desktop. There were barelly any other OS, but only few systems which had just few percent of market.
There is no point to have 10x better exchange alternative(just an examplen), when canonical has to let users use slow and buggy evolution or not so flexible webfrontend as user interface for exchange alternative. Outlook is not perfect, but outlook is working good and has relatively consistent UI.
People are lazy to learn and they learn only when they must. So ubuntu, and linux in general, needs few more optimizations to tackle microsoft in business world at this point.
Linux did come a long way and is a great system, but it is far from easy, especially when you get into trouble that you have to do something which is not covered by standard use of PC. For business world, i’ll just name few microsoft applications, which are important: Outlook for mail, Powerpoint for presentations, Visio for diagrams and MS Project for project management. Yes, there are linux alternatives to these products, which are good and even very good, but none excels beyond above listed applications. I know there are better mail clients than Outlook, but there is no single one which is tightly intergrated with mail servers, like outlook with exchange. Impress is not bad, but does not reach functionality offered by Powerpoint. And there are definetly better products than MS Project, but or they are standalone products, or they don’t have flexible interface (usually webui), etc.
There is one other issue that linux have when we are talking about market share. This is at same time one advantage as it is an issue. And this is that linux has many tools for same problem solution. There are like 10 different diagraming tools available, but each and every one is not as good as microsoft product. We need to unite and to try to force as few options as possible.
And last point is: “win the user at home and he will request same working environment at work”. Whats the point to force linux and/or ubuntu on users, when they can’t use it easily at home? Yes, everyone and even my grandma can install linux… almost. Because this is solved relatively good. All these installation wizards provide enough for people to install it easily, but there are still problems in that. Just think what happens when they try to install Ubuntu on some hardware, which is not that usual (my own example: Abit IP 35-E, cheap intel p35 board with ICH9 southbridge, a crippled version of full supported ICH9R. This stopped me from installing Ubuntu on my main PC so i had to opt for Vista on it.), they will look at it and say… crap… can’t use it, forget it.
Linux and Ubuntu got a looong way ahead, but there is still a road to be conquered. We had to get users used to Linux first, before we force them server applications down their throats.
Alecs
The main problem with focusing on the server market will allways be : a professional need a rock solid server distro,
and desktop users won’t like it.
How many desktops are running Debian Stable ?
How many professional servers are running Debian SID ?
I think they need to work on finding a niche for server edition. Instead of just adding a lot of ISV support that Redhat and SLES already have, why not find some ISVs that aren’t currently working with Linux and engage them to get it working on Ubuntu?
Microsoft’s backoffice raid of the 90′s was aided by three things, IMHO: (1) the desktop was already ubiquitous, (2) The server carried the same familiar interface that the desktop had (note that at the time the actual core of the OS was different — NT vs DOS), and (3) they offered things the Unix servers of the day did not.
In the case of Ubuntu server, (1) The desktop version is popular among linux users, but by no means ubiquitous, (2) The server version carries none of the user-friendliness of the desktop version with it — it’s old-school Linux administration, even if you install a GUI on it, (3) What does it offer that RedHat and SLES don’t already offer?
That’s all the armchair quarterbacking I’m up to this morning.
The first one thats smart enough to catch a hold on the desktop will win the server world as well.
Are you selling servers to admins or to business people? Most of the time at least in my experience business people get to weigh in way to heavily on these decisions. What do they go with? They go with what they HEAR about the most either through marketing or personal experience. The hear and see MS all the time. You get Ubuntu to be well know on the desktop, laptop and netbook and they’ll be more likely to investigate it when someone talks about using it on a server. Thats how Ubuntu will get into the business driven shops rather than just the tech driven shops.
If you forget the desktop now you can just bout forget the whole thing.
Also I agree with Alan above and thought along the same lines. What does CLI Ubuntu have thats going to sell it over CLI RHEL, CentOS, Debian or any other command line installation. Yes they can add some different tools…but this is FOSS…everyone else will have it too sooner or later. Focus on the enterprise desktop and management of those desktops if you really want the money. Help one of the FOSS messaging projects build a complete alternative to Office/Exchange/Outlook. The battle is on the desktop.
I think desktop Linux got them were they are, so it should not be overlooked. The main reason Windows Server gained popularity was because it had a bunch of desktop users that were willing to give it a try.
Other Linux companies focused on servers and, in my opinion, hampered the proliferation of Linux into the world. Windows servers were not used because they were better, they were used because they were familiar.
Okay here’s why the readers are right.
When Microsoft launched it’s server products in the 1990s Microsoft was already ubiquitous with business desktop PCs. Microsoft was trying to network these PCs together with it’s own proprietary networking standards and protocols that nobody else wanted to adopt. So it made sense for Microsoft to push it’s server products.
Canonicals primary user base isn’t businesses and Canonical aren’t pushing anything unique like a queer networking protocol. So far as I can see, it’s home users that make up the core of Ubuntu users. It’s Linux newbies. It’s the user that wants something not Microsoft but thinks he/she can’t afford a Mac.
Canonical have done a great job working with major tear one OEMs and getting Ubuntu preloaded onto PCs and netbooks. It would be criminal not to leverage this opportunity. OEMs like Dell and HP clearly aren’t interested in building their own OS. This could be a way for Canonical to monetize their Linux and OSS business model and at the same time differentiate it from Red Hat and Novell. They can do this and still steadily build their server presence.
Canonical should be focusing on home users, small and medium businesses then the big server market. No matter how dependent we become on the web. We’ll still need a secure desktop OS to take us there.
Aikiwolfie: You’re surfing across our sites faster than the site owners are today
I respect your opinion, but stay tuned. Big server news coming. Ubuntu Server Edition is nearing a tipping point.
There are a few simple reasons why I disagree with you:
SUSE Linux Enterprise Server
Red Hat Enterprise Linux
and all the rest.
Linux already does servers, and it does them very well!!
What it doesn’t to super well is desktops.
I think I will put my faith with Mark and his team, he is the man with a plan.
I think I can see part of it, use the desktop success to push the brand and gain trust, make money on the sever side eventually.
Redhat OWN the server side atm and you cant fight them head-on and win.
Windows server became popular to a large extent because of people being familiar with the desktop product.
Canonical are trying to commoditise the Sever side and eat Windows server alive – an ambitious goal;)
Also don’t forget that the Desktop brings with it a huge Community and Mind share.
One last thought, a large scale success on the desktop helps get hardware manufactures to write drivers for Ubuntu. I still can not believe that Dell now sells PC’s with Ubuntu installed! Big Win.
Lets not forget Ubuntu Bug#1
Bug #1 (liberation): Microsoft has a majority market share [edit]
Quote:
Microsoft has a majority market share in the new desktop PC marketplace.
This is a bug, which Ubuntu is designed to fix.
Non-free software is holding back innovation in the IT industry, restricting access to IT to a small part of the world’s population and limiting the ability of software developers to reach their full potential, globally. This bug is widely evident in the PC industry.
Steps to repeat:
1. Visit a local PC store.
What happens:
2. Observe that a majority of PCs for sale have non-free software pre-installed.
3. Observe very few PCs with Ubuntu and free software pre-installed.
What should happen:
1. A majority of the PCs for sale should include only free software like Ubuntu.
2. Ubuntu should be marketed in a way such that its amazing features and benefits would be apparent and known by all.
3. The system shall become more and more user friendly as time passes.
https://launchpad.net/ubuntu/+bug/1
I hear there’s a new Linux server OS coming out soon…I can’t quite remember it’s name. Deb something. Debian. 5.0. Oh, yeah, that’s right. Okay. Debian 5.0 is coming out soon.
So why does Ubuntu need a Server side, now?
Nobody Important: Stay tuned. I will see your Debian 5.0 and raise you with Ubuntu Server Edition’s major news in a few days.
I am a little confused here as to the parallels, or the lack thereof, drawn between the results of the survey and the direction of Canonical with respect to Ubuntu. I think They are doing a great job for the users as reflected in the survey, but for their shorter-term money goals that guard against bankruptcy. Server and Desktop markets segments are separate. In fact there are sub-markets withing those segments. Each sub-market requires individual treatment, like a samba PDC at a small office vs. a server farm. Should Ubuntu emulate RedHat and SuSe? The problem I see with Linux (U R right) is, where it meets business, not enough is being done. We are all hoping to make money, and hopefully do some good, out of F/OSS in its various forms of engineering, software, business, etc, but that’s a philosophical journey that we are all taking. Windows was all about integration and they invested a great deal giving what their users wanted. As proponents of Linux we need to delete the edges and polish the faces presented to users. We benefit from multiple distributions, remixes, variants, etc where we can deliver solutions. Microsoft does not have this, and also lacks the ability to deliver goodwill in the way Linux can. Some time ago I was reading about two market substitution/complements models, for example real networks server (market 1) makes more money in market 1 because the buyers have confidence in the ubiquity of the free player (market 2), than if both market were not free. Piracy is not as prevalent in the US, but consider the levels of piracy outside the US. For all the variety of reasons piracy is supported out there, there is a benefit in using Linux instead. That benefit can be monetized. Inexpensive distribution? Protection against piracy litigation? The warm feeling in your heart that you got something good for free and its legal? Then people can ultimately pay to support social benefits by opening up custom applications built for businesses. Ultimately intellectual property is protected and shared and turned into beneficial profit.
I have to agree with the author over against the readers. Look, the server room is already the place where Linux generally is earning footprint in the marketplace. Redhat and Novell have both built a business model on Linux, and have done well, though some purists have been put off by what they see as Novell’s Mephistophelian pact with Microsoft.
But Ubuntu does stand to penetrate the server space right alongside Redhat and Novell. Canonical has been very savvy in building incremental alliances with some not-insignificant names including IBM and HP.
I would not in any way denigrate the progress in the desktop space, but it will come more slowly – technology adoption takes a fairly predictable adoption pattern that follows an S-curve – Linux has not reached anywhere near the tipping point in the desktop space, but in many organizations is already becoming disruptive in the server space. Many companies who are completely uninterested in Linux as a platform have Linux based technologies either in their server rooms or as embedded systems. The latter technology I think is also poised to penetrate the marketplace for Linux more rapidly than in the desktop space, IMHO.
Look at the big names which have already certified Ubuntu server – Lotus Notes and DB2 do not have majority market share, to be sure, but neither are they insignificant players.
There’s much more to be said, but in conclusion I’m inclined to agree with Joe that the server space is lower hanging fruit for Canonical than the desktop space.
What has Canonical brought to the Linux Server world that Redhat or Novell hasn’t?
Ubuntu’s strengths are community and the desktop and they should play to those first and foremost, when… not if, some of those strengths result into major innovation on the server end then they would be a force to be recognised in the server world.
However I believe that too much focus on the server too early will ultimately harm the community and ground breaking innovation that’s happening there and should ubuntu fail on the desktop, I don’t see how they would hope to succeed on the server.
The server business is well covered by Linux already unless I am mistaken, but more effort in that direction is always good. I feel what should be tempered is the continuous release of newly numbered and named versions of Ubuntu and other Linux distributions. I would like to see some long-term stability in them. Microsoft had problems with Vista but is continually upgrading and learning from it since its release.
I feel stability is extremely important and is not all a matter of perception. Any Linux newcomer who researches Ubuntu hesitates (as I did) at downing,burning and using a buggy,only-supported-for-18-months version of (place name here) when they can visit the site and down Arguing Aardvark a few months down the road? Stay put and make the next version of Ubuntu last. Iron out all the problems instead of racing ahead only to fall behind. Engender some commitment from potential new users by committing to them with some long-term stability.
This just seems to me like more confusing “Ubuntu” and “Linux.”
The needs you have for a server distro are about the polar opposite of what you want in a desktop distro — even assuming stability is paramount in both (when even that tends to be a different compromise).
Ubuntu would be my last choice for server distro because it makes just about all the wrong choices — overaggressive adoption of new technology, not enough stability, etc.
The Microsoft example is the exception that proves the rule. I don’t think Microsoft’s server OS has had that much of an impact on their desktop deployment. They have basically two OSes — and many people deploy BSD / Linux over Windows because they feel the server OS doesn’t fit their needs.