Slashdot Mirror
Linux: One quarter of the server market by 2003
weezer writes "LinuxToday has a nice little article about a Dataquest study that claims that Linux will account for about 24% of the server market by 2003. "
← Back to Stories (view on slashdot.org)
From what I've read, it looks like they went with a hybrid that consists of the old-style NT domains and some aspects of the "regular" directory structure, as known from, say, Novell's NDS:
m parisons/ADandNDScomp.asp
you have your regular tree structure with a Root, Organization, Organizational Units and leaf objects, but you also have a Forest, which is supposed to be a collection of separate directory trees that trust each other and are able to synchronize and replicate directory data through some sort of trust relationship....I played with it a bit and then went back to NDS 8, which is infinitely better in every aspect - scalability, stability, ease of use, not to mention maturity (it's been in the works since mid-90's).
As for LDAP, M$ AD has an LDAP connector, although it is "implemented a bit differently" (read: "embrace and extend").
If you are into FUD, here's what M$ has to say about their "new" invention in regards to NDS:
http://www.microsoft.com/windows/server/eval/co
The rebuttal from Novell can be found here:
http://www.novell.com/products/nds/nds-rc.html
Hope this helps.
Does anyone else find these numbers a little strange? 24% by cost = 14% by volume?? The competition here is Linux, NT, commercial Unix in various flavours, proprietory OSes and free Unixen. All but the free alternatives cost more than Linux (assuming you pay, which is not actually necessary), so how come the volume percentage is lower than the cost percentage? What am I missing? The term 'Server Applience Revenue'. The article makes a distinction between appliences and 'traditional servers'. Also, the prediction is for the cost of hardware + software sold in 2003. It appears that they are expecting the higher end applience boxes to be running linux. Later in the article, they refer to 'traditional server'. Dataquest noted that Linux will account for much less in the traditional server market. By 2003, it will represent $1.9 billion, or 3.4 percent of traditional server revenue. That represents 450,000 units, or 8.1 percent of shipments.
I feel like picking a fight with everyone who thinks they are right. - Rainmakers
Our network admin (for the entire enterprise, not just a "site admin") is a certified ignoramus. She's in charge of NT and Novell (with a couple Unix, etc oddballs) machines and she knows diddly. Calls the programmers in to create batch files, has never heard of Perl, still creates NT users with the GUI (even though 99% of them are identical and could be scripted), etc ad nauseum. What's even funnier is that she's gone to week-long training classes on NT, Exchange, SQL Server, etc and yet still knows nothing. (The only reason I said the above was "funny" was that I don't have to work with/for her. If I did, I would quit after a month.)
---
Put Hemos through English 101!
Linux MAPI Server!
http://www.openone.com/software/MailOne/
(Exchange Migration HOWTO coming soon)
Dataquest isn't really a group I consider a purveyor of reliable information. Just look at their past predictions on the growth of NT (They've been wrong a LOT). For a long time they were considered a mouthpiece for MS products.
It's not that this is not good news for Linux users -- it's definately a sign that the winds of public opinion are changing direction. It's just that Dataquest reports tend to be reflective of what is trendy and buzzworthy, not necessarily what is statistically sound.
Eviscerati.Org: All Hail the Eviscerati
These articles all have misleading titles. If you read the actual contents of the articles, they explicitly say:
... Dataquest noted that Linux will account for much less in the traditional server market. By 2003, it will represent $1.9 billion, or 3.4 percent of traditional server revenue. That represents 450,000 units, or 8.1 percent of shipments.
It estimates Linux servers will account for $3.8 billion, or 24 percent, of the worldwide server appliance revenue by 2003. By volume, that represents 14 percent, or 1.1 million units.
A server appliance is not a "traditional" server. What operating systems are used in server appliances? Probably *BSD, Linux, NT, and a tiny fraction of proprietary others? So in four years, which operating systems are used in the other 76% of server applicances? Which operating systems are used in the other 92% of "traditional" servers?
cpeterso
Does anyone else find these numbers a little strange? 24% by cost = 14% by volume?? The competition here is Linux, NT, commercial Unix in various flavours, proprietory OSes and free Unixen.
All but the free alternatives cost more than Linux (assuming you pay, which is not actually necessary), so how come the volume percentage is lower than the cost percentage? What am I missing?
If you read the C|NET article, you will find that the study refers to the "server appliance" market, not the "server market". This refers to out-of-the-box server "solutions" like the NetWinder.
The article also says that Dataquest predicts Linux-based systems will account for 8.1 percent of the "traditional server" market. I think that's underrating Linux by quite a bit.
I think that data is off by a little bit.
First, I think Linux already accounts for at least 20% of the server market -- it's just not openly shared. It's the kind of thing that engineers sneak past dim-witted IT managers.
Second,
Linux, developed in 1991 by Linus Torvalds, then a student at the University of Helsinki in Finland, is considered a more stable version of the older Unix operating system.
Shouldn't that be, " More Stable than Windows NT ? " Because Linux is DEFINATELY more stable than NT, but I thought it was pretty much "On Par" with the other Unices.
And Finally,
" Linux will continue to grow as an operating system deployed on new server hardware, but much of the hype surrounding Linux will fade once Microsoft ships a service pack for Windows 2000 server, " Brown said.
Let's not forget that the so-called " Linux-Hype " survived the release of Windows 98®, so I'm pretty sure the Hype of Windows2K isn't enough to smother the well deserved credit and attention that Linux receives.
Reasoning: More NT "admins" couldn't handle a command line to save their lives. Sure, it may be faster and more powerful and, on occassion, even a more simple way to handle things (compare setting up Apache vs. IIS), but just the thought of doing anything beyond clicking on radio boxes will scare the hell out of 'em. Sort of an advanced technophobia -- an extension of the people who are terrified to even use a computer lest they screw something up.
Sad, really. There are a lot of NT admins (not all, and not even the majority, but still a lot) who should feel guilty about picking up their paychecks. I mean, if a company pays you to come up with computing solutions, it seems really unfair to rule out an entire line of products and possibilities just because you don't know enough to really do your job.
Of course, this article was about appliance servers, but we'll ignore that for now.
----
Every year during my review, I just pray the words "slashdot.org" aren't mentioned.
Unless CNET was in a hurry to get some copy out, this isn't much of a report. 25% of server-appliance sales (not servers, not server sales, just preconfigured colorful boxes, e.g. the netwinder). I took more issue with the "traditional server sales" paragraph, such as it was:
"...Linux will account for much less in the traditional server market... 3.4% of traditional server revenue." Well, duh -- Linux servers aren't sold as such, they're sold as blank machines and one installs Linux from elsewhere. Take the 3-5 million servers kicking Linux around out there (assume they represent the same growth rate as between now and DQ's 2003; probably wrong, but skip that for a moment). DataQuest's 450k servers over a $1.9B market share give a server price average of $4222; if this is 3.4% of total revenue, then the total market is $55.83B. Back to the 3-5 million servers (suppose 4); at $4222 each one finds $16.88B, or 30.2%.
All that said, the 4 million could be way off, and server-sales figures are a miserable estimate of relative product performance except where marketroids are concerned.
http://www.sunworld.com/swol-04-1999/swol-04-idcli nux.html
IDC said that the Linux server market was 17.2% in 1998 and predicted an annual growth rate of 25% until 2003. That means that by 2002, Linux would be more than 40% of the server market.
Given that Linux makes A LOT more sense than NT or commercial Unixes for 'server appliances' (no per-seat license fees, no duplication license fees, and open source adaptability), I think Linux will be a considerably larger part of the market than 23% by 2003.
I, for one, expect Windows 2000 to flop rather nicely in the server market. Here's why:
....)
1. Anyone who has been paying attention to the past few Microsoft releases would know that they have been increasingly less popular than expectations predicted. IIRC, MS shipped about half as many Windows 98 upgrades as they expected in the first few months; 98 became dominant not on the strength of upgrades but on the strength of the growth of the new-PC market. The idea that people need to upgrade just for the sake of upgrading is declining.
2. NT itself has peaked and is in decline in some server markets, notably the Web server market. Microsoft Web servers have been declining in market share for months on the Netcraft survey of Web servers. People are realizing that Microsoft systems are not reliable and scalable, much less enterprise-ready. Why buy more of the same?
3. W2k will break some third-party software that runs fine under current NT releases. This is just how MS operates. Sites which expect not to have catastrophic failures will wait and test W2k for some time before deploying it as a replacement for current NT systems, as the Gartner Group recommended several months ago. In the meantime, sites which rush ahead and move to W2k will have the usual early-adopter problems that any new system has. This will generate horror stories which will reduce other sites' interest in W2k. Vendors of Unix- and Linux-based systems will, if they know what they're doing, capitalize on these failures. (They may even FUD Microsoft back, though of course I wouldn't support that
4. W2k will not live up to the hype. MS has already (quietly) admitted that the widely-hyped Active Directory is not a directory at all (in the sense of X.500 and LDAP) but rather a flat pseudo-directory. Expect more of the same.
5. Finally, there's Y2k. Nobody's going to make any major changes to their mission-critical now, and in January there will be enough mopping-up to do that they'll delay still more.
Windows 2000 may well be a good deal better than NT. But will people gamble their businesses on it to the degree they've gambled them on NT? I think we can expect not.
Meanwhile, Linux-based systems are doing nothing but growing. What's the safe bet if you want to be using a platform that will continue to grow, survive, and thrive over the next ten years?
Of course, in fact they rely solely on "dumb luck" in order to be right about anything.