Not only is Linux having trouble breaking market share for desktops, server shipments actually grew at a smaller rate than Windows server shipments. “Somewhat surprisingly, Linux seems to be running out of steam a little. After nearly four years of double-digit revenue growth, the Linux server sub-market accounted for only USD1.5 billion in sales in the second quarter of 2006, an increase of only 6.1 percent. IDC didn’t say this, but it could be that the mainframe market has saturated itself with Linux and is no longer consuming Linux MIPS like it has for the past several years. Linux server shipments grew 9.7 percent in the quarter, and most of the revenue and shipments were on X86 and X64 servers. Windows-based server shipments increased by 11 percent.”
Where to start and how to start…
First the title of the article is “The Server Market Struggles for Growth in Q2, Says IDC”
Sounds more like windows is needing to be replaced yet linux isn’t? Okay, I am guessing!
It is a very general article that covers servers in general…I wouldn’t run out and get more MS stock tomorrow or anything.
How many of those who see a single quarter’s slowing of growth in Linux server SALES (which say little about Linux deployments) as evidence that “Linux is running out of steam” would see the same in a single quarter’s stagnation or reduction of Windows server sales?
The question I think is more relevant is how it was counted; was it ‘server shipments’ or ‘operating system shipments’ – for me, I prefer to use the statistics of ‘operating system shipments’ as many choose to go through xyz vendor, but purchase the support contract directly from the Linux vendor itself.
I can’t say that there is ‘slow growth’ given that the sales for both Red Hat and Novell are going gang busters are the moment. With that being said, Novell has a long battle to win in that they’re investing large portions to make their distribution and ecosystem more competitive, whilst at the same time, raising the profile of their brand so that Linux is no longer synonymous with Red Hat.
As for the over all server shipment; Sun’s sales have been going incredibly well; Solaris shipments are up, OpenSolaris is pushing forward.
Ultimately, I think the greater emphasise should be put on this; a customer who moves to Solaris or Linux, is a customer who isn’t going to be using Windows anytime soon – its a win for all when a customer moves in the direction of *NIX, irrespective of what camp you place yourself into.
Well, if Linux grows slower than Windows, and Windows is 3 times the sales of Linux, Linux has zero chance of catching Windows.
Well, if Linux grows slower than Windows, and Windows is 3 times the sales of Linux, Linux has zero chance of catching Windows.
What part of “over one, (1), uno, ein, un, een quarter” don’t you understand?
Not nearly enough time to indicate a trend.
Edited 2006-09-01 01:38
kinda off topic..yet not.. I work in a restaurant where we use a POS system for our orders. We had a problem with the system that in the end had 3 different techs looking at the system. Being the in house resident geek and cook/chef, I asked them if they had any experience with the linux based Vu-Touch system. None of them did. None of them even had any experience on Linux desktops or servers. None of them even knew of any Linux Distros beyond Red Hat. Maybe they were just idiots.. but I ended up being the one to fix the system. It was a BIOS issue. These guys barely Knew windows. They would have been utterly in the weeds doing anything on a nix system. That being the case.. I can understand why some techies and admins arent making the effort to install Linux servers.. they have no clue, and don’t wish to learn.
or they have learned, that with windows a reboot fixes the problem, until the next reboot, but with unix, if its broken, its time to break out the kung-fu…
Many distros have tools to do away with the “kung-fu”, but it all boils to the same thing. Some of us learn from our mistakes, and others get jobs administering Windows.
Edited 2006-09-01 01:21
If by “kung-fu” you mean learning what needs to be done and fixing it right the first time then I will take kung-fu any day of the week.
I would rather be a kung-fu lackey than a re-boot king.
– Quality
– Speed
– Ease of use, and management.
– Security ( steal from OpenBSD ).
Those should be the linux goals.
Well, speed right now is already done. Unless Vista comes out with some trick, the latest Suse is faster. I work 20 hours a day and keep coming back to it for certain tasks. Quality is a very complex subject and I wouldn’t comment. Ease of use: that is the real opening window for Linux. The GUI paradigm is not always the smartest one. Just try this; gnuplot versus Excel for graphics; if you have more than one gnuplot can be used with great efficiency and EASE.
The big problem now is that every piece of software is available in Windows: Lyx, Emacs, Gnuplot, Latex (WinEdt runs circles around Kile), etc..
It will be VERY tough to compete with Vista in software availability unless MS really screws it.
Edited 2006-09-01 02:27
Windows is better at management ATM. Good remote/multiple machine managing software, WMI, standard interfaces (and coherent documentation!), opposed to a “free market” race that is currently present on linux; in fact each distro has their own schemes for configuration (consolidation here is needed, urgently!). Besides, MS invests heavily in marketing and “training” for their systems.
To get things done for linux, though, it’s often just enough to convince few that there is a need for improvement in specific area. So once the problem is recognized, things just GET DONE (certainly there isn’t a lack of manpower in open source world).
Linux needs more quality apps to appeal to a broader user base and to stop copying Windows, I mean who wants to be reminded of Windows anyway
Argh not again. When we refuse to do things the Windows way, people complain it’s not like Windows. When we do ’em the Windows way, people complain we’re ripping off Windows.
Maybe we should just have a big sign at the border of Linuxland that says: Windows users ->
Look what a heap of shit Winlinux was? Linux should be unique not a clone of MS stuff. After doing point and click on XP for so long it was a joy to telnet and use SSH to install PHP, I had forgotten the attraction of a console screen, reminds me more of star trek than anything on the Windows.
Windows remote management better than Linux? Who are you kidding?!
What Linux really needs is an OEM supporter. Joe-end user is looking to buy a computer then use it for what he needs to do. He can do that right now with an Apple, or a Dell, and he gets the support he expects when things foul up (even if that support isn’t free). That’s the deal regular end users are missing if they try to install Linux for themselves (this is a silly idea, no regular user will do this).
As a bonus, an OEM can ship proprietary products (like video codecs, and flash) that are actually illegal right now on Linux, because an OEM could negotiate the appropriate contracts.
So there you have it, problem solved. If we want more linux desktop market share, you need an OEM to start shipping computers with it, and it has to be good enough at end user tasks to beat Mac OS X and Windows Vista, out of the box, without the need to install third party applications – which are currently missing anyway – a problem that might be helped if there was an OEM backing a particular consistant stable Linux configuration (Platform). ISVs like stable well supported platform APIs.
I’d totally buy a nice quiet SFF Dell Alienware system running Ubuntu if it had all the proprietary stuff (flash/codecs/etc.) installed by default (or at least very easy to set up when I turn it on the first time – like automatic).
If Playstation 3’s Linux is open enough, it could be that OEM (I have no faith in Sony’s ability to get this right though ). It’d be easier to get an Adobe to port Shockwave Player to this platform if there are hundreds of thousand of users (millions?). I would imagine that they’d also be more willing to port the same software to other Linux hardware platforms, since it’d probably be easier to port from Linux on Cell to Linux on x86, than it would from Windows to Linux (maybe not though). (Of course it’d be even easier if it was written for a hardware agnostic platform like .NET/Mono or Java! )
And yes, all of this means getting along with proprietary vendors better. IMHO OSS makes the most sense for commodity software, like Operating Systems. The innovation will always be in proprietary software – the missing process right now is how to turn successful (which maybe means the new defacto standard) software into OSS without loosing an important business advantage.
I guess maybe the answer is that there needs to be a business calculation – at what point does it become more expensive to maintain a large code base (like say the Windows kernel) than it is to work cooperatively with your competitors to maintain that same code base (thereby sharing the burden)? When Macromedia (now Adobe) switched Flex off of the proprietary Dreamweaver code base and on to the open source Eclipse code base, they seem to have made this calculation (and in this case chose another vendor’s code base over there own – neatly sidestepping the “not invented here” mentality that some firms suffer from ).
Edited 2006-09-01 03:43
It’s very, very dangerous to think we should compromise with proprietary software vendors or DRM supporters at this stage. OSS is at the tipping point; if proprietary vendors see FOSS software producers backing down, they will be all over them like a pack of hungry wolves.
This is about server shipments. not Joe User.
We deployed about 50 Linux servers (Suse, Debian) this year. I am sure ALL these servers where not counted as the article says because they only count servers with pre-installed server Os’s. Since Windows mostly comes pre-installed i think the server figures of Linux must be much higher. Also people (linux admins) have find their way to download Linux and use online docs en tech help from forums/wiki and irc, so no need for Redhat, Suse etc. they just bought the hardware and deploy their own server.
Edited 2006-09-01 05:42
I would have to agree wholeheartedly. I work for a hosting company that is a partner with RedHat. We put 100+ servers online daily running Linux. Of course none of these get counted because they’re loaded via kickstart.
I just setup an Edubuntu server utilizing LTSP for a lab here at work for the small school we have on campus. That server would not have been included, nor the Ubuntu server that I put in place last week for our Proxy/Web Filter.
If this shows anything, it’s showing that Linux is maturing to the point where Windows Admins (such as myself) do not have to rely on vendors to hold our hands while we learn Linux, because Linux has become so much easier to maintain and configure.
How were the sales of servers with NetBSD doing? Are they keeping pace?
KIDDING!
More fictional numbers taken out of thin air from our friends at IDC. Does anyone listen to them anymore?
Hell, yeah. I mean, people still listen to Yankee Group, and last I heard they hadn’t even sacked Didiot.
Most gnu/linux distros allows sysadmins to simply update their installations to be up-to-date even for critical pieces (kernel, glibc…)
With microsoft if you want to be up-to-date you have to pay.
GNU/Linux is selling less than before because it WORKS and it LASTS!
A better investment indeed
Windows servers are so good that all ISP’s use them, cheap too!
It’s a crime to use Windows as a server platform if you depent on the service…
Do you remember the time when hotmail.com and microsoft.com were running on BSD?
And what about that UNISYS Unix to Windows transition ‘try’? That was some good fun.
It just a lot less pain to explain a tie-wearer (aka. “manager”) you need 20 Colored-Clicky-Wizard-Windows Server instead of the Unknown-Geeky-Linux stuff you dont really have support for.
Another wave of pain touched us, when they wanted to migrate the whole infrastructure of a stock exchange to Windows (because the CEO went golfing with Ballmer, no joke). What the hell did they smoke?
So choosing Windows as a service platform has, for me, always something to do with lack of knowledge about computer systems. But thats just my 2 cents.
This report is off. why does it appear that linux has less new servers? Because we decided not to use Redhat or
Suse and decided to use Debian and gentoo!!!!!!!!!!
Sometimes beancounters are idiots.
idiots, well,
every that does please linux fan, are idiots.
That’s a wonderful philosophy, when application are badly design, no need to make it better, because users are idiots. Unstable, no need to make it better, users still are idiots. Distro X have an isued, users are idiots because he doesn’t use distro Y that lingeek #1 recommended. apps Z on other system have betters reviews, no need to get competive, because reviewers are idiots.
When I used the apache configuration gui provided on my fedora core 5 system, well it simply doesn’t work. So I configured myself. Then after I installed lighttd, the startup config gui, still included, simple crashed. I had to configure it by hand. I Still dont understand how unstable application can be ship with a system. They should be carefully test to give a stable , efficient as well as pleasant systeme to work with. (my soundblaster wasn’t working out of the box – quite unpleasant in 2006)
I know I didn’t use _your disro V plus the application wasn’t stable, so I must conclude that I am an idiot because of reason F. But still, I can’t wait to see the article “why 2008 is the year of linux desktop”.
When I used the apache configuration gui provided on my fedora core 5 system, well it simply doesn’t work.
Well, of course that proves beyond a shadow of a doubt that Linux is borked. After all, nothing in Windows ever breaks.[/sarcasm]
Imagine if IDC were into Linguistics, we’d get posts like “Article in Broken English On Popular OS-centric Website Proves English a Non-Starter”.
Microsoft increasing sales by putting their customers gonads in the upgrade vice and turning the handle does not equate to linux servers being converted to windows.
I never really like these sorts of server shipment reports that implicitly suggest that all servers bought in the world a) come with a commercial OS pre-installed and b) that the pre-installed OS is never wiped once the server is shipped.
If you go to Dell’s UK or US site, you can buy PowerEdges with no operating system installed – how does this report account for those? And what about finding out what’s done with such a no-OS server when it goes into production use? My company, for instance, buys Dell server with no OS and puts a free Linux on it (usually a Red Hat variant) and I suspect that this kind of action accounts for more servers than all the pre-installed-with-commercial-Linux servers out there.
Any report that mentions Linux server market share and fails to research the number of *actual* Linux servers that go into production use isn’t worth the keyboard it was typed on.