A recent IDC report showed Linux servers continuing to increase market share for x86 architecture with a second consecutive quarter of double-digit growth, but the bigger news could be Microsoft’s even bigger surge with Server 2003. The software giant’s Server 2003 showed modest gains in Q1, with IDC reporting that Microsoft Windows Server revenue was USD 4.8 billion in Q1. This number represents 10.4% year-over-year growth and a gain of 1.9 points of revenue market share over the same period in 2006. Windows encompassed 38.8% of all server revenue in Q1 of 2007.
Since it’s more expensive to run Microsoft servers, market share based on revenue numbers will show an artificially large Microsoft number. If Windows has 38.8% of all server revenue, it follows that a far smaller fraction of servers are running Microsoft software (especially considering that CentOS has a revenue share number of zero, no matter how many copies are being run).
Yes, that’s just another example of how difficult it is to get an accurate number when dealing with market share (servers and/or desktops).
No doubt this will be touted by some as proof that Microsoft is “dominating” the server marketplace, and that Linux is “past its prime”, even though these numbers do not actually reflect actual market share…
It’s simple. Linux market share is proportional to Microsoft’s desperation.
That sort of comment fits well with my top-5 list of the most ridiculous OSNews comments.
I know it’s rather stupid now that I read it again. I apologize. But it’s more ridiculous that it was given a +5. :S
The EU uses these same sorts of stats to determine that Microsoft is indeed “dominating” the server market, and bases its punitive measures on such stats. Microsoft bashers normally support such punitive measures.
So, either these stats are meaningless, misleading, or whatever negative adjective you want to use, and thus there’s no need for the EU’s punitive measures regarding the server space, or the stats do have meaning and Microsoft is indeed increasing its server share. Which is it?
Do you enjoy putting words in people’s mouths?
One can support punitive measures against Microsoft for anti-competitive tactics (or simply because they broke the Law), but until you give me a precise example of someone who specifically claimed that a) MS is “dominating” the Server Market and should be punished for this (good luck in finding this, btw), and b) that these numbers do not support domination of the Server Market by MS, then you might have an example of hypocrisy on your hands. Otherwise, this means absolutely nothing.
Personally, I don’t know of *any* pro-Linux advocate who will claim that MS dominates the Server market (because, well, it *doesn’t*). Perhaps you can provides some links for us?
Edited 2007-06-11 20:11
even though these numbers do not actually reflect actual market share…
Who really cares about “actual market share”?
Let me put it another way: would you rather have, for example, 1% of market and make some money.. or would you rather have 99% and make $0?
Microsoft is in business of making profit, not claiming highest market share and no revenue. Apple is another example.
Edited 2007-06-11 18:00
Exactly. Revenues is a common way to build statistics and abstract units from actual meaning. It is no doubt that Linux probably has a larger installation base if measured as units, but it’s also no doubt that a rather significative part of such installatons are test machines, campus machines where students install their own web servers, a lot of Apache installation which only show default page and so on.
They can be significative in terms of units but they count 0 in terms of business meaning. So measuring revenues you actually measure how many companies are using your system to make some money, not your cousin’s website for fun.
Unix users could easily say that in terms of units or installations they lead and there is no doubt about this (even if, given the number of W2003 growth, for how long?).
Linux growth and Windows2003 growth means Linux is cannibalizing other Unices until now, while Windows is eating Unices’ marketshare. The real problem would be if Linux will be the final challenger, having replaced most other Unices. Until the fight in Unix field is not over, we cannot know which is the real business value of the Linux platform. For example, free Solaris could easily reverse the trend if Sun plays its card well.
I agree with you mostly about the business meaning part.
The business install base is what’s more important in the big picture.
There is one more factor that needs to be considered though – how many of the WS2003 servers are replacements for WS2k servers? A Linux machine can still run fine after all these years… but a WS2k machine is now without official support since August of last year (IIRC). So… it wouldn’t be a big surprise to have people moved to WS2003 because of this.
Well, of course a few units will be upgrades but as the Microsoft share is growing very quickly and because WinNT/W2000 didn’t have such large share, I guess we can say Windows2003 is significantly increasing MS share in server market.
Anyway, there are also many WinNT and Windows2000 still in place. This was a major problem for Microsoft as many customers didn’t want to upgrade to Win2003.
Anyway, Unix share is still larger than Microsoft’s but if the current pace holds, it won’t last forever.
Anyway, Unix share is still larger than Microsoft’s but if the current pace holds, it won’t last forever.
Hmm… are you sure about that?
http://seattlepi.nwsource.com/business/308645_msfteu23.html
Microsoft’s gains in server market share ‘worry’ EU
European Competition Commissioner Neelie Kroes said it’s worrying that Microsoft Corp. continues to boost its position in the server market three years after the European Union fined the company a record $664 million.
…….
At the start of the commission’s antitrust investigation in 1999, the agency said Microsoft’s share of the worldwide work-group server market was between 35 percent and 40 percent. That share rose to about 60 percent between 2001 and 2003. The current share is about 75 percent, the commission said.
That only represent a portion of the server market.
I agree. So, how does he know for sure that Unix has bigger server market share than Windows? He does not. Could be the opposite and I suspect it is because of Windows dominant position on the desktop — it tends to lead to Windows-only networks, where both desktops and servers are running Windows. I am not saying it is always the case (sure, it is not) but it tends to be.
And just one more thing:
That only represent a portion of the server market.
..but it is significant enough for EU to start going after Microsoft, right? So, I’d say it represents major portion of that (server) market. (Web servers are not the most numerous servers in any company, except those that provide that [web hosting] service, etc).
The EU isn’t going after MS because it has a dominant position on the Server Market. That was MollyC’s assertion, but the truth is more complex than that.
That requires so many leaps of logic, it’s like being inside the mind of a Microsoft marketroid.
The EU isn’t going after MS because it has a dominant position on the Server Market. That was MollyC’s assertion, but the truth is more complex than that.
Definitely more complex, but Microsoft will bow now because they cannot afford to loose EU market. And how highly publicized “big migrations” showed, Linux is not an alternative good enough. I wouldn’t be surprised if EU winks to Solaris now that it is open.
Actually, it *is* a good alternative, which is why “big migrations” as you call them continue to happen.
BTW, the FUD about Munich is getting old. They had problems at the start (being trailblazers of sorts), but things are much smoother now, and the Munich migration can be said to be success (despite Microsoft’s propaganda machine portraying it as a failure).
Actually I read at least 3 cases where projected “big migrations” turned out to be more difficult if not stopping at all. I’m almost sure I read about them on this website so some searching could help.
The best chance to Linux is in emerging markets. Which are not small markets at all! But there, you can try building the ecosystem from the group up.
But that doesn’t mean Linux is better than Microsoft. For example, Windows has been kicked out of China (expecially in public administration) but that was not for quality matters rather on political ones.
(and if I have to tell you the truth, I agree with chinese govt. Go figure… same applies to Europe as well)
I can think of only one (London), and that hadn’t actually gone underway. Most likely, Microsoft made them an offer they couldn’t refuse.
As far as I know, the other large-scale migrations – though clearly not simple undertakings by any means – have all kept going.
*Any* mass-scale migration is going to be difficult. You’re right, though, it’s not simply a matter of quality, but rather of not being locked in with a single vendor. That’s both a political choice and a pragmatic one – IT has become too big a portion of our lives to entrust it to an aggressive, ambitious monopolist.
I do agree that Linux does great in emerging markets, though again you’ve got to consider that Microsoft tolerates a great deal of piracy in order to make sure they keep a competitive edge there as well.
I can think of only one (London), and that hadn’t actually gone underway. Most likely, Microsoft made them an offer they couldn’t refuse.
Yes this was a case of carrying out a Linux trial to deliberately induce Microsoft to make them a very generous offer.
*Revenue* share…
How much does Win2003 cost compared to Win2K?
Free of charge and unregistered Linux servers do affect the market, in fact.
At least some of the people running those servers would have been Microsoft’s customers, if there was no Linux.
Free Linux servers are decreasing revenue for Microsoft and others. For Linux vendors, too, maybe even more than
for Microsoft.
Obviously, measuring revenue doesn’t measure market share (although I’d think that companies like Red Hat care more about the former than the latter, while the reverse would be true for CentOS).
That’s why the important stat isn’t the absolute numbers, but the relative increase/decrease of the numbers. I think it’s common knowledge that Windows Server 2003 is increasing its share steadily, regardless of what the absolute number is.
“I think it’s common knowledge that Windows Server 2003 is increasing its share steadily, regardless of what the absolute number is.”
Why? I didn’t know it, so its not that common.Its 2007 is Windows Server 2003 replacing earlier incantations of itself or Unix servers. The article actually states it thinks that Market share is not being taken away from Linux. When we talk about revenue is Microsoft squeezing more money out of its customers? We know there is a growth in x86 due to a minor revolution on the CPU. What percentage is the Windows growth on that? Is that the thing thats driving the Windows revenue growth by giving Microsoft houses a real reason to upgrade? How about the weak dollar?
Thats ignoring How successful are commercial Linux Vendors against Free alternatives…are they offering enough over then to compete successfully? This year.
The article is practically meaningless from anything but what it is….a revenue perspective, but other than X86 driving hardware sales it offerers no insight.
The trouble I find is that its hard for *anyone* on here to get a real perspective of whats going on . Its a global marketplace. This is not product A is better than product B this is a different. Statements like
“The vast majority of the time, they buy machines preconfigured with an OS that meeets their needs and then buy support at the same time that they purchase their machines”
is simply untrue, as I can guarantee than *everyone* here, has little to no knowledge of how;why and what the company down the road is using, never mind globally.
I picked the statement out because its simply a lie, but I don’t think its intentional, companies are pretty isolated in what they do, thats why we have these reports come out.
http://news.netcraft.com/archives/2005/12/05/strong_growth_for_debi…
You can see the fastest server growth for at least for web servers is Debian.
Edited 2007-06-11 19:20
Since it’s more expensive to run Microsoft servers, market share based on revenue numbers will show an artificially large Microsoft number. If Windows has 38.8% of all server revenue, it follows that a far smaller fraction of servers are running Microsoft software (especially considering that CentOS has a revenue share number of zero, no matter how many copies are being run).
Except that’s not the way it works out most of the time. Most organizations don’t just buy servers and then download an OS & pave them with a completely different operating system. The vast majority of the time, they buy machines preconfigured with an OS that meeets their needs and then buy support at the same time that they purchase their machines. Sure, there are maverick companies that don’t follow these general principles, but there aren’t many.
I’m sure you have data to back those assertions up…right?
I’d get off my high horse there if I were you. Take a stroll back through your commentary and you’ll see time and time again you yourself try to impress upon others your view and experiences without any “data to back up those assertions”.
The way I am reading his his comment is that the cost of ownership and acquisition (which is a component of ownership in any event) is more than just downloading and installing CentOS or any other Linux distribution. There are many more factors to include.
This is a point that so many anti-Microsoft people time over selectively forget about. Linux is in no way “free” in terms of dollar costs for the overwhelmingly vast majority of organisations. It requires specialised knowledge, experience and training. This costs an organisation. There’s a learning curve which also has a cost attached to it.
Paced against Windows Server for instance, yes there are costs associated with ownership also, this is a fact, however one would have a valid point in my view, that it’s less in some particular areas, if only because through the client OS (2000, XP, etc) we’ve become rather accustomed to how Windows operates, where various parts of the OS exist and so forth.
People will try and drive the hype machine as why organisations are choosing Windows Server over Linux (e.g. Microsoft’s advertising is just so darn good and convincing when really it’s all lies), but in reality these organisations will have done their homework, built cases for both alternatives and made a decision with one, to achieve their own objectives and strategic goals.
If you so some basic research you’ll see that many reject the lower TCO of Linux in favor of lower TCO from Windows Server, and to that end it would be rather silly of any of us who were not involved in that decision making process to suggest that we’ve got sufficient grounds to ridicule such decisions.
Of coarse, I don’t have any “data to back up those assertions” off hand. It’s just my experience and point of view from what I have read both on and offline.
Perhaps *you* should get off your high horse. Tomcat stated that there aren’t many companies who install OSes on servers. All I did was ask more details about this rather bold statement.
First, what does “not many” mean? 25%? 10%? 5%?
Second, how does he manage to know what most companies do? Has he done surveys on this? Read studies? I’d like to have that information too, because I find it interesting.
This has *nothing* to do with being pro- or anti-Microsoft. Rather, it’s about staying away from vague declarations that can’t be verified in a rational debate about the subject at hand.
That happens, but you *also* have people who aren’t as diligent in planning their IT strategies, and who just go with the “big name” because, well, that’s what they know. The fact is, we can’t really say what the proportion of each, and therefore it’s useless (and fallacious) to pretend that one case is more prevalent than the other.
You make it sound as if there’s no incompetence in the corporate world, especially as it pertains to IT. The one thing this tells me is that your experience in this matter appears quite limited…
Come on, virtually *all* of those studies were funded or commissioned by Microsoft. I could have told you what they were going to say before they came out…
If you so some basic research you’ll see that many reject the lower TCO of Linux in favor of lower TCO from Windows Server, and to that end it would be rather silly of any of us who were not involved in that decision making process to suggest that we’ve got sufficient grounds to ridicule such decisions.
I’ve been, directly and indirectly, involved in several processes deciding between Windows and Linux servers, and I’ve never ever seen anybody chose Windows because of TCO. In fact I can’t ever really recall TCO being an issue.
Most organizations don’t just buy servers and then download an OS & pave them with a completely different operating system.
depends what the server will be used for. I (and most people in field that I know) have done just that countless times in several companies for all kinds of projects. Sure not the main production server, but for internal stuff, development, skunkwork projects etc. that’s normally the way it works.
This is my experience too – internal webservers and the like are usually done this way – buy a box (or put one together from bits lying around), put an OS (linux usually) and Apache on it. Can often be done out of petty cash, no revenue for anyone. I’ve even got people who phone me for “per-incident” help on these servers. None of this, I should think, gets on the revenue figures for Linux
at work we have 230 servers running; about 40 of them w2k/w2k3 and amongst the unix servers, a few run vmware esx with 30+ virtual RHELs….
At my previous work, there also was a shift towards unix(alikes).
Yes, the revenue stuff is a specific part that counts and is pretty much msleading to what really is being run.
Dude, “msleading” is like the best description of Microsoft marketing I’ve ever seen. Kudos for that!
With all of the Microsoft-funded studies they do, I seriously doubt any of their reports hold merit anymore with IT departments.
The one thing I would absolutely LOVE to see is a complete, totally objective report from a team that isn’t funded by Microsoft or any Linux-based company.
Now considering that this isn’t really an option at this point in time, I would more than welcome a report paid for by two parties at the same time: Microsoft and Red Hat, with each paying an equal share for the report so that there isn’t favoritism towards the company who may have paid more, along with an independent oversight committee to make sure the rules are fair and everyone is playing by them without any deviation of said rules.
Of course, that really isn’t an option, either, is it?
Regarding the comments about the number of servers vs. amount of resources, that seems to be threatening to boil over into an argument…
Note that any correlations you might make about the absolute number share cannot be drawn off of numbers about the revenue shares. For one, although you can get fairly (not perfectly) reliable numbers for Windows, you can’t make the same assumptions for Linux; what is an expensive solution with expensive support from one vendor is a fire-and-forget distro with forums (if that) from another, yet they still come under the heading of ‘Linux.’
It may seem safe to say in general terms that more revenue means more people using it, but that may be skewed by the number of large sites running one or the other — for example, Myspace is a purely Windows-served site, and despite the high revenue that obviously generates, the number of Myspace servers does not come to even a thousanth of a percent of the market. Nevertheless, that one statistical spike will sharply push the revenue stats pro-Microsoft.
I’m not saying anything either way pro or against either, only that revenue numbers can’t give usage numbers readily. You can’t get there from here.
Exactly.
Wow. We agree on something. I better take the rest of the day off and get some rest, I must have a fever or something… 😉
I’ll be much better off, if they give me the raw numbers and I’ll make my own conclusion.
Dell will sell you servers with no operating system installed (even ones where they don’t even offer a preinstalled Linux only Windows). Where are all these OS/less machines from Dell and the other hardware companies in IDC’s survey. The vast majority of these machines will have Linux or an open source Unix installed on them by the customer not Windows.
These figures could really alter the result of the survey. Think of all the servers out there running CentOS, Fedora core, Debian, Ubuntu, BSD’s and OpenSolaris. They are probably running on machines purchased without an installed OS.
Dell will sell you servers with no operating system installed (even ones where they don’t even offer a preinstalled Linux only Windows). Where are all these OS/less machines from Dell and the other hardware companies in IDC’s survey. The vast majority of these machines will have Linux or an open source Unix installed on them by the customer not Windows.
Not necessarily so. Some larger organizations purchase volume licensing agreements which often include both OS + applications, and they don’t want to pay for the same operating system twice.
Who cares? The article states that Windows Server as 38.8% of the server market while Linux has 12.7%. So that means Windows already won but both are increasing in usage.
All that should matter is how much marketshare an OS has in your home.
Linux, FBSD, and OpenSolaris account for 95% of the marketshare in my office.
What about you?
All these worlds are yours except Europe.
All these worlds are yours except Europe
LOL. I havn’t heard that in a long time…
Reminds me of how much I love that book.
My take – Many people were using Novell Netware and traditional UNIX at one stage. A percentage of them moved to Windows NT4. Some held on. They waited until Windows 2000 Server was released and then jumped ship – by then it was clear that UNIX was dying, and that Windows could do pretty much everything a traditional UNIX box could do, without the proprietary hardware and expensive licenses. It was also clear that Novell was dying in the market. Linux was nowhere to be seen at a corporate level imho.
Move forward in time until 2005 – Linux has become more well known and used in the corporate environment. Windows 2000 is getting very old in the market. From a performance point of view, 2000 is not as good as a modern Linux server from that period of time (security/reliability, features, hardware economy, etc). Along comes Windows 2003 – it’s a lot better server option from Microsoft. It’s more reliable, more secure and more aimed at the server market than previous Microsoft options. It sells like hotcakes. Why?
*MOST* workplaces will NOT just download a Linux ISO and burn to disk and put it on a server. They WANT support. Let’s look at commercial support options for Linux – Redhat, Novell & Suse. All of them very expensive, in fact it’s cheaper to get support from Sun Microsystems for Solaris 10 on the Intel architecture. Microsoft? Not only is the software cheaper (per seat), but the support is cheaper. And many would argue, better. Let’s look at Redhat’s offerings:
https://www.redhat.com/apps/store/server/rhel.html
and
https://www.redhat.com/apps/store/server/rhelap.html
Prices in US dollars, and not cheap I might add.
Now let’s also consider several other things – there’s a lot of open source 3rd party tools that are free that bridge any gap between what Microsoft Windows offers, and what a Linux variant offers.
Then there’s also the cost of hiring someone – hiring someone to handle your Linux server data migration, setting it up, securing it, keeping it running reliable etc isn’t cheap. You’re going to find many ex UNIX sysadmins, who are out of jobs because of the UNIX decline are migrating to Linux due to its many similarities to the traditional UNIX. They demand more money than a Windows sysadmin as a traditional rule.
All of this adds up to money, serious money. Microsoft Windows wins out here, because it really is the cheaper option. I think you’re really, and honestly, going to find very few workplaces downloading Debian and using it on a server.
Let’s also take into account that most IT managers in large to small firms will not trust open source software, they consider it a toy, too unreliable for their needs.
My experiences in the work place over the past 5 years have shown that none of them have employed Linux on their servers. None. Read it again, NONE. And, none of those employers were considering moving to Linux. In fact, they all upgraded to Windows 2003 Server (from earlier versions of Windows server, mainly 2000). You can mock what I’ve said, trying to disprove it, but as good as Linux is, many corporate environments won’t touch it for a variety of reasons, and this translates into sales for Microsoft. Windows 2003 has been a very good product for Microsoft, I’d argue their best in many key areas. Ask any Windows sysadmin and they’ll all say the same thing – Server 2003 is the best so far from Microsoft and that’s winning a lot of friends.
Linux fans – you really need to understand that Microsoft is learning from its mistakes, and in some key areas trying to retain market share by making a better product. Whilst Microsoft is still engaging in a lot of FUD, it is also improving its products and people are noticing it, and staying with Microsoft.
In all honesty, Linux was far too slow out of the starting box. It should have been where it is now on the desktop and server way back in 1998/1999. There’s too little, too late imho. I’m not knocking Linux, it’s a fantastic operating system, it’s just that it doesn’t suit most people’s needs for a variety of reasons, some of that obviously to do with Microsoft’s illegal monopolistic lock-in behaviour. I’m not saying Microsoft’s perfect, or all rosy, so don’t get me wrong. For most networks, most people I know in the roles are picking Windows 2003, it does what they want, how they want it and is a big step up from previous Microsoft offerings, at a decent price. That’s what wins customers and customers translate into sales/profit.
Dave
“Let’s also take into account that most IT managers in large to small firms will not trust open source software, they consider it a toy, too unreliable for their needs.
My experiences in the work place over the past 5 years have shown that none of them have employed Linux on their servers. None. Read it again, NONE”
How stupid do you have to be to not run a heterogeneous environment. I’m not even advocating Linux. That really is insane. After reading that there is little point at looking at your other points.
Edited 2007-06-12 10:36
“Not only is the software cheaper (per seat), but the support is cheaper. And many would argue, better. Let’s look at Redhat’s offerings: ”
This is dishonest for at least two reasons:
1.You completely neglect to mention some of the biggest cost factor in Microsoft solutions; Client Access Licenses and Terminal Server Access licenses.
The Red Hat server offerings allows you to use an unlimited number of clients to access each server. The Microsoft server offerings do NOT. Look at the pricing here: http://www.microsoft.com/windowsserver2003/howtobuy/licensing/prici….
It adds up pretty quickly. Including 25 clients, it costs USD 3999,-, a fair bit more than than the prices quoted by Red Hat. Also note that the prices quoted for Windows Server only includes business hours support, thus it is not fair to compare these prices with Red Hat’s 24×7 support options, but rather the 12×5 ones.
2. The Red Hat server support pricing includes support for an enourmous amount of server software that Windows 2003 pricing do not.
Ouch! What a bozo!
I am seeing more and more Linux servers where I work.
The place used to be a Unix place, but some
dumb-ass CEO asked for Windows servers.
I work as a consultant, and I see the same
Linux growth at my clients companies.
And those Linux servers are taken care of
by regular IT staff.
Not old Unix gurus, sometimes youngters fresh
out of school.
Today’s average 14 years old nerds can
administrate a Linux box, it’s no longer a secret art.
And no, IT staff are not necesseraly buying support contracts.
By the way, have you ever contacted
Microsoft’s support. It’s horrible. Most of the
time, you end up with nothing useful and a
huge bill. Community-based help is way more useful.
Yay, so Microsoft has fastest growing revenue share, so we should all be using NOTHING BUT Microsoft products!
In other news, Ruby is the fastest growing programming language, so we’re using it to write drivers for embedded systems, W00t!!
These market stats always seem to overlook the fact that linux-based appliances are displacing conventional servers, both Win and *nix based, in datacenters of all sizes. Particularly in the security and storage space, you’re finding applications migrating to purpose-built appliances (that are really optimized servers) running some flavor of linux, whereas in the past these applications would have run exclusively on standard servers.
Companies like Nortel and Avaya are migrating or have migrated their IP PBX platforms from embedded Windows-based to linux based. We’re not talking about small potatoes here in terms of revenue generated, but the emphasis in appliances is on the application or service, rather than the platform. Often times the admins aren’t even exposed to the underlying platform, but I don’t think that’s reason enough to dismiss it.
I won’t bicker about revenue-based market share stats, but just pointing out that the deployment of network services and applications is evolving and linux is probably much better positioned to capture that transition than Windows is. But frankly there’s more than enough room for both.