Linked by Eugenia Loli on Sat 30th Sep 2006 00:12 UTC
Talk, Rumors, X Versus Y Microsoft Windows will not suffer irreparable damage on the server side at the hands of the Linux operating system over the next five years, Gartner analyst George Weiss told attendees at the Gartner Open Source Summit. In fact, in terms of worldwide server operating system revenue, Linux would come in below both Windows and Unix by 2011 in spite of its enormous growth, he told attendees in a session entitled "Enterprise Linux: Has it Arrived?"
Order by: Score:
Over the next five years?!?!?!
by Milo_Hoffman on Sat 30th Sep 2006 00:43 UTC
Milo_Hoffman
Member since:
2005-07-06

Wow, nothing earth shaking will happen in 5 years...way to GO OUT ON A LIMB THERE....

jezz...people pay for this crap?

Reply Score: 5

RE: Over the next five years?!?!?!
by Babi Asu on Sat 30th Sep 2006 14:21 UTC in reply to "Over the next five years?!?!?!"
Babi Asu Member since:
2006-02-11

Why believe Gartner? Gartner didn't know that since 1997, every year is Year of Linux Desktop.

Reply Score: 5

Bah, we dont need to be #1 anyway
by Xaero_Vincent on Sat 30th Sep 2006 00:47 UTC
Xaero_Vincent
Member since:
2006-08-18

Being dominate means all the attackers would target thier viruses, trojans, worms, etc. on our platform. While Linux is great at fighting them off, its still not a perfect system. Malicious progs could exploit bugs in vulnerable elements like the kernel, apache, mail, xserver, so on. Distributions not protected by advanced security measures, such as SELinux, AppArmor, GRSec, would get the worst beating.

Linux is not going anywhere when Vista ships, even if it boasts major new security features that rival Linux and what not.

We'll just take some concepts from Vista and use them to improve Linux.

Reply Score: 1

bytecoder Member since:
2005-11-27

Since when do "viruses, trojans, worms, etc." target servers?

Reply Score: 1

Mathman Member since:
2005-07-08

Would it not depend on the virus, trojan or worm in question? Some of them have certainly targeted servers.

Reply Score: 5

dylansmrjones Member since:
2005-10-02

Right now, actually. If not directly than indirectly. How do you think so many servers have been brought down? By random malfunctioning?

Reply Score: 3

Xaero_Vincent Member since:
2006-08-18

An unprotected Linux box is just as vulnerable as Windows; remember old versions of Lindows/Linspire? Thankfully openSUSE has many easy-to-use security features to help protect my system. :-)

Now I just need to find a good anti-virus program for Linux and set up "sudo".

Reply Score: 1

dylansmrjones Member since:
2005-10-02

A standard Linux system is not unprotected.

Lindows/Linspire was deliberately insecure.

Reply Score: 5

Coxy Member since:
2006-07-01

Why?

Reply Score: 1

dylansmrjones Member since:
2005-10-02

Well, Lindows/Linspire was meant to resemble Windows ;)

Reply Score: 1

Sphinx Member since:
2005-07-09

Since the epoch, but to date I believe the only successful attack was a mail worm quite a long, long time ago.

Reply Score: 1

Sort of to be expected
by Jody on Sat 30th Sep 2006 00:54 UTC
Jody
Member since:
2005-06-30

I have to agree with the guy above, there "analists" get paid pretty facking good money spit out what amounts to common sense.

One of his statements is interesting though. He said Linux will sell below UNIX by 2011 and this has certianly no been the trend so far. I believe Linux has been eating UNIX's server market share at a fairly steady rate and actually passed it about a year ago.

I know Sun has been doing better lately, but if he predicts UNIX will make a comeback I am guessing he expects the progress to continue.

Reply Score: 3

RE: Sort of to be expected
by kaiwai on Sat 30th Sep 2006 03:40 UTC in reply to "Sort of to be expected"
kaiwai Member since:
2005-07-06

I know Sun has been doing better lately, but if he predicts UNIX will make a comeback I am guessing he expects the progress to continue.

The death of UNIX was based on whether or not Solaris actually picked up some traction and moved forward - these analysts are privy to sales information which most of us can't obtain, and specifically, who the big customers are (which dictate whether its just a bunch of enthusiasts purchasing or actually high profile, big accounts).

Solaris is now straddling x86 and SPARC; worse case scenario, Intel and AMD multi-core make SPARC systems irrelevant (doubtful), Sun is still in a good position to generate revenue off their x86 side of the business.

With that being said, analysts also predicted the dying of UNIX back when Windows NT was released, they also predicted massive growth and the replacement of x86 servers with Itanium. The simple fact is, none of that has happened.

Itanium is stuck in a niche, high end UNIX RISC machines are being replaced, where practical, with low cost x86 machines running some sort of UNIX based operating system, and high end UNIX boxes are being used to replace mainframes as mainframe features are scaled back into UNIX based machines.

Reply Score: 3

RE: Sort of to be expected
by NotParker on Sat 30th Sep 2006 05:22 UTC in reply to "Sort of to be expected"
NotParker Member since:
2006-06-01

I believe Linux has been eating UNIX's server market share at a fairly steady rate and actually passed it about a year ago.

No. Both Unix and WIndows each have about 3x the revenue as Linux on servers.

And Linux growth is petering out.

"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 $1.5 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.

In terms of platforms, the Unix server market declined by 1.6 percent to $4.3 billion and unit shipments of Unix boxes declined by 1.8 percent to 166,000 units.

The Windows platform, which has been jockeying for dominance with Unix for the past several years in the server market, accounted for $4.2 billion in sales in the second quarter, according to IDC, and increase of 3.1 percent. Windows-based server shipments increased by 11 percent."

http://www.itjungle.com/tlb/tlb082906-story01.html

Edited 2006-09-30 05:23

Reply Score: 4

RE[2]: Sort of to be expected
by chemical_scum on Sat 30th Sep 2006 10:22 UTC in reply to "RE: Sort of to be expected"
chemical_scum Member since:
2005-11-02

And Linux growth is petering out.

the Linux server sub-market accounted for only $1.5 billion in sales in the second quarter of 2006, an increase of only 6.1 percent

The Windows platform ... accounted for $4.2 billion in sales in the second quarter, according to IDC, and increase of 3.1 percent.

So Linux revenue growth is twice that for Windows servers. Petering out my foot.

Also the survey does not cover servers sold without OS and yes you can buy them from Dell. I am sure that there are an awful lot of these sold to have Debian and CentOS etc or for that matter a BSD installed on them, particularily for web servers and "edge" of the datacentre functions.

Reply Score: 2

RE[3]: Sort of to be expected
by NotParker on Sat 30th Sep 2006 14:51 UTC in reply to "RE[2]: Sort of to be expected"
NotParker Member since:
2006-06-01

"So Linux revenue growth is twice that for Windows servers. Petering out my foot."

1) Linux revenue growth used to be 20,30 and even 40%. So yes, I would say it is petering out.

Linux: 6.1% of 1.5bil = 91.5mil
Windows: 3.1% of 4.2bil = 130.2mil


2) Windows unit sales growth is higher than Linux: 11% to 9.7%. That hasn't been true for years.

"Also the survey does not cover servers sold without OS and yes you can buy them from Dell."

IDC tries to account for all situations.

For example, our K12 distict has bought over 200+ Dell servers without an OS on them over the last 5 years because we get the Academic discount with our software reseller.

In fact, many Enterprise shops don't buy the OS from Dell. They install from discs they get throught their enterprise agreements.

Edited 2006-09-30 14:52

Reply Score: 4

RE[4]: Sort of to be expected
by chemical_scum on Sat 30th Sep 2006 15:51 UTC in reply to "RE[2]: Sort of to be expected"
chemical_scum Member since:
2005-11-02

Linux revenue growth used to be 20,30 and even 40%. So yes, I would say it is petering out.

I can't ever remember linux revenue growth being in the 40% range. But over the last couple of years or so when Linux growth rates were in 20-30% range Windows revenue growth was in the 10-15% range. That Linux has about twice the revenue growth has remained constant what has declined or "petered out" is overall growth in server sales revenues.

IDC tries to account for all situations.

For example, our K12 distict has bought over 200+ Dell servers without an OS on them over the last 5 years because we get the Academic discount with our software reseller.

In fact, many Enterprise shops don't buy the OS from Dell. They install from discs they get throught their enterprise agreements.


So how exactly does UDC take in to account your situation and the sorts of situations I pointed out in my original post? I have never seen this explained anywhere.

Windows unit sales growth is higher than Linux: 11% to 9.7%. That hasn't been true for years.

What has happened here is that more of the low end of Linux server sales has shifted onto local installs of free as in beer Linux distributions on OSless boxes. That is how Windows can have half the revenue growth of Linux sales but about the same unit growth in sales.

Reply Score: 2

RE[5]: Sort of to be expected
by NotParker on Sat 30th Sep 2006 22:58 UTC in reply to "RE[4]: Sort of to be expected"
NotParker Member since:
2006-06-01

I can't ever remember linux revenue growth being in the 40% range.

I would say 63% in 2003 to 42.6% in 2004 to 20.8% in 2005 to 6% in 2006 is definitely petering out.

http://news.com.com/Linux+server+sales+show+high-end+trend/2100-734...

"Linux server revenue grew 63 percent to $960 million from the fourth quarter of 2002 to the fourth quarter of 2003, IDC said."

http://www.infoworld.com/article/04/11/24/HNlinuxserversales_1.html

"Quarterly sales of servers running the Linux operating system topped $1 billion for the first time during the third quarter of 2004, analyst company IDC reported Wednesday. With year-over-year revenue from Linux server sales up 42.6 percent..."

http://www.idc.com/getdoc.jsp?containerId=prUS20074406

"Linux servers generated $1.6 billion in quarterly revenue, the fourteenth consecutive quarter of double-digit growth, with year-over-year revenue growth of 20.8%."

Edited 2006-09-30 23:02

Reply Score: 1

RE[6]: Sort of to be expected
by raver31 on Sat 30th Sep 2006 23:13 UTC in reply to "RE[5]: Sort of to be expected"
raver31 Member since:
2005-07-06

even though the percentage numbers are coming down, the amount of sales are increasing.

I see you know nothing of economic growth.

Reply Score: 2

RE[7]: Sort of to be expected
by NotParker on Sun 1st Oct 2006 00:21 UTC in reply to "RE[6]: Sort of to be expected"
NotParker Member since:
2006-06-01

"even though the percentage numbers are coming down, the amount of sales are increasing."

2006 - 6.1% of 1.5 billion = 93 million annual increase

2003 - 63% to 960 million = 371 million annual increase

Annual growth in Linux sales has dropped 278 million - a 75% drop.

"I see you know nothing of economic growth."

More than you it appears.

Reply Score: 2

RE[8]: Sort of to be expected
by Ookaze on Mon 2nd Oct 2006 11:57 UTC in reply to "RE[7]: Sort of to be expected"
Ookaze Member since:
2005-11-14

"even though the percentage numbers are coming down, the amount of sales are increasing."
2006 - 6.1% of 1.5 billion = 93 million annual increase
2003 - 63% to 960 million = 371 million annual increase
Annual growth in Linux sales has dropped 278 million - a 75% drop.
"I see you know nothing of economic growth."
More than you it appears.


As long as the growth is positive, that means they sold more than before. This is basic maths !
Now, of course it will peter out, how could it ever accelerating the growth ? it's reaching a plateau in growth rate, which is to be expected.
A drop in annual growth is a very twisted way for trying to say a negative thing from a positive thing (this 75 % loss of growth rate over 3 years you talk about doesn't mean anything). Peter out is another attempt to make a negative view on this. The growth rate is the acceleration, not even the speed.
They're not even talking sales here, but REVENUE !
Now, another basic math for you :
1.5 billion + 93 million > 960 million, so the amount of sales revenue is increasing.

Reply Score: 1

Solaris is stronger than people realize...
by tomcat on Fri 29th Sep 2006 20:07 UTC
tomcat
Member since:
2006-01-06

It's been enterprise-ready for a long time, it's pretty darned bulletproof, supports hot-swapping and quite a few technical innovations, and it's going to continue to grow its market share now that Sun announced its intention to open source OpenSolaris. Still, this isn't bad news for Linux advocates: Linux is still growing its install base significantly. And, if MS stumbles with its server products as it did with Vista, then Linux and the traditional UNIX brands could steal some market share from Microsoft.

Reply Score: 4

Xaero_Vincent Member since:
2006-08-18

I liked Nexenta, which is the closest I'd ever come to Solaris. But its missing a feature I demand: Accelerated drivers.

I hear the Nvidia drivers are better on Solaris than FreeBSD. Unlike FreeBSD, SLI is supported and 64 bit drivers are available. However FreeBSD has better ATI support with DRI and Linux has the proprietary driver.

Maybe I'll go with a Nvidia card in my next system.

Reply Score: 1

kaiwai Member since:
2005-07-06

IIRC DRI is being ported to OpenSolaris; its going to take some time unfortunately ;)

Reply Score: 1

Rayz Member since:
2006-06-24

And, if MS stumbles with its server products as it did with Vista, then Linux and the traditional UNIX brands could steal some market share from Microsoft.

... But what if they don't stumble?

Reply Score: 1

raver31 Member since:
2005-07-06

But what if they don't stumble?

Is the pope a catholic ?

Reply Score: 1

deb2006 Member since:
2006-06-26

Nope - cannot second that. In traditional companies you see that Solaris is being replaced by Linux. This is a decision that was taken some time ago. The companies won't suddenly change their minds just because of OpenSolaris.

SUN has quite a few technical gadgets. Some of those are really, really good. Take zfs, e.g. However, Linux e.g. has other mechanisms that might take a bit longer but that produce the same result (lvm, xfs, e.g.).

If you ask me: SUN has - over the past few years - taken many steps in many directions. OSS is just one direction. And that's simply not enough to convince people. So it's about technology, but it's more about reliability.

Edited 2006-10-01 02:08

Reply Score: 1

his opinion
by backdoc on Sat 30th Sep 2006 01:41 UTC
backdoc
Member since:
2006-01-14

OK. He gets paid to go out on a limb and commit to something. I can respect that. But, I don't think he's been using the same operating system that I have. 'cause, what I'm using is great, and it ain't Windows.

Reply Score: 3

Lets get real
by SEJeff on Sat 30th Sep 2006 03:05 UTC
SEJeff
Member since:
2005-11-05

One of the standards "Enterprise" IT shops use is the top 500 super computer list. I guess this analyst failed to notice how Linux completely owns it compared to Unix or Windows:
http://www.top500.org/stats/27/osfam/

Nothing but Linux scales as well horizontally for as cheaply as it does with the level of customizability you are afforded. This is the reason why time and time again, massive clusters and supercomputers are running (surprise) Linux.

If Vista has all of the security improvements that Microsoft is touting such as ASLR, processor no execute (NX), etc, it will be interesting. For once, Linux and Windows will be close to similar regarding security. This will increase innovation in areas of manageability and system provisioning. This only benefits the consumers and that is us.

From my personal opinion, this report is totally bogus and this guy is a Microsoft fanboi. On the otherhand, my job is a Linux systems admin so obviously, I'm biased.

Also notice how this guy says Unix server Operating System revenue is above that by several million dollars. Maybe that is because most of the large shops compile their own version of Linux for FREE while buying thousands of Unix (HP-UX/Solaris/AIX) license costs millions. Does that make what he says correct? Yes. Does that makes what he says very deceptive and super biased? Yes. nuff said.

Edited 2006-09-30 03:09

Reply Score: 5

RE: Lets get real
by Robert Escue on Sat 30th Sep 2006 04:19 UTC in reply to "Lets get real"
Robert Escue Member since:
2005-07-08

You mind telling us how many "Enterprise IT" shops use supercomputers to run database, document management, LMS/LCMS, CRM and other business applications? This is truly comparing apples to oranges! Just because Linux "owns" the supercomputer world doesn't mean it owns the server rooms of all companies big and small. And for all that supposed superior scalability, I don't see people trashing their Windows, Solaris, AIX or HP-UX installations for it. In the last week I have read two articles about companies dumping Linux for Solaris, so what does that mean, that maybe George Weiss might have it right?

And how does Vista, a client operating system fall into a comparison of server operating systems? It is yet to be seen whether SuSE makes a dent in the Desktop OS market, provided Novell stays solvent long enough for companies to see that it is a product worthy of their atttention (and money).

And you obviously don't know squat about OS licensing when it comes to Solaris/AIX/HP-UX (fortunately, I do). In the purchases of hardware and software for the projects I have worked on since 2001 (including Navy/Marine Corps Intranet or NMCI), the cost of the applications far outstripped the cost of the OS. In many cases the OS licenses were "thrown in" to sweeten the deal for management to select a particluar platform. The SPARC hardware that EDS purchased for one site alone was in the millions of dollars, so you can bet Sun threw EDS a bone. If you look at the cost of a Solaris 10 license, you will find that it is free for individual or commercial use. And even when Sun charged for Solaris 9, a multiple CPU license costs $300.00.

Now let's take a look at RedHat, we have several hundred RHEL installations at where I work. In order to apply patches to those machines is going to cost us around $100,000.00 because of RedHat's "subscription" nonsense. Now with Solaris, AIX, HP-UX and Windows I can easily deploy patches and it costs me nothing.

Now compare that with Microsoft and their pricing for Enterprise products (Exchange/SharePoint/SMS/ISA Server, etc.) and it comes down to what your shop can support based on product experience and skill level, what the workforce uses and what they are willing to pay for it.

And which "large shops" spend their time compiling Linux besides Google? It can be done, and I'm sure the few places that do have a full time staff to do it. Somewhat unrealistic in the grand scheme of things, either you pay for support to the various vendors or you spend it on your own support staff but either way your going to spend the money.

Reply Score: 5

RE[2]: Lets get real
by kaiwai on Sat 30th Sep 2006 05:41 UTC in reply to "RE: Lets get real"
kaiwai Member since:
2005-07-06

I think Meatloaf put it best in the song, "you took the words right out of my mouth", in regards to that post.

Having used Windows at an ISP with IIRC 30,000 or so end users, I can assure you, we had no problems keeping it up; Windows 2000, reliable, Windows 2003, very reliable.

When I hear the doom and gloom merchants from the ABM group, I brings to question whether they have the necessary IT skills for that job in the first place, if they can't setup a system correctly.

Reply Score: 2

RE[2]: Lets get real
by SEJeff on Sat 30th Sep 2006 14:42 UTC in reply to "RE: Lets get real"
SEJeff Member since:
2005-11-05

I work for ticketmaster... We have thousands of Linux servers around the world in a massive grid. This includes web, email, database, etc. All of them are versions of Linux similar to CentOS compiled and managed by a homebrew configuration management system similar to cfengine.

Please explain how we could save money using anything other than Linux? OpenSolaris would be the only other choice but it's not our inhouse expertise. BSD is lacking in really good SMP support that scales.

Reply Score: 1

RE[3]: Lets get real
by Robert Escue on Sat 30th Sep 2006 16:13 UTC in reply to "RE[2]: Lets get real"
Robert Escue Member since:
2005-07-08

Considering the amount of money Ticketmaster brings in, you would think they could afford a lot more than using Linux in that fashion. The labor costs must be through the roof, but if it works, great.

I just don't buy into the idea of a complex IT solution on the cheap. At some point you are going to spend the same money, whether it is on hardware, software and support, or personnel to keep your "free" solution going. I work for a number of people who think F/OSS software is great because they don't have to pay for it, and think nothing of making us jump through hoops to make it work until they get in trouble for the solution failing. The scenario I just painted actually happened at where I work. The end result is they are now going to use a commercial product instead of the F/OSS solution.

Free or low cost does not necessarily mean better or easier to manage, or more effective. It just means low cost or free, which is appealing to management, particularly the "penny pinching" kind. If you have to custom tailor it to your environment, you either hire the staff to do it or buy the products that come with the functionality "out of the box". That is the difference between a Windows, Solaris, AIX, HP-UX and a "roll your own" Linux solution. You choose the tools based on what is needed and what skills you have to support it.

Reply Score: 1

RE[4]: Lets get real
by Ringheims Auto on Sat 30th Sep 2006 17:37 UTC in reply to "RE[3]: Lets get real"
Ringheims Auto Member since:
2005-07-23

Well, I'm sure you can find proprietary software on wich you must spend loads to get working too. These kinds of unixsoftwares tend to be geared towards professionals in any case, be it on GNU/Linux or on Solaris, and are not ment to be easy to use and to set up.

And keep this in mind: "Free" in Free Software stands for "Freedom", not "cheap" or "gratis".

Why spend much money and time on software wich you CAN't own...?

Reply Score: 2

RE[5]: Lets get real
by Robert Escue on Sat 30th Sep 2006 22:14 UTC in reply to "RE[4]: Lets get real"
Robert Escue Member since:
2005-07-08

The difference between F/OSS and proprietary is standards compliance, support and functionality. Most commercial products it is a matter of configuring the features you want and extending the rest through some development effort. Some of the F/OSS products we use has poor documentation and support, which significantly increases the administrative effort and cost of a "free" product. If it was my call, I would use commercial software because it provides what management wants with minimal headaches. And despite its high initial cost (depending on the product used), it invariably ends up costing less over the lifetime of the project. If a F/OSS product can meet the requirements and provide the necessary functionality, cool.

We use a mix of both commercial products and F/OSS software and being that I work at a US Government installation, software ownership is not a big deal. Management could care less about RMS and "free software", what they care about is something that works and doesn't cost a lot of money. Unfortunately something that is selected based on flimsy requirements (or none at all) and cost alone does not always make for a good hardware/software solution. The example I used in the previous comment ended up costing 8 figures before management "threw in the towel" and chose to use a commercial product, most of that was labor costs for the development team.

I see that at least one person doesn't agree with me (previous comment modded down) but that is the way it is at where I work. I believe in promoting the "right tool for the job" and that tool might not always be F/OSS software. Promoting F/OSS software to the exclusion of everything else does no good for anyone, especially if it doesn't meet the requirements of the project and gives people the wrong impression of F/OSS.

Reply Score: 1

RE[6]: Lets get real
by Ringheims Auto on Sat 30th Sep 2006 22:42 UTC in reply to "RE[5]: Lets get real"
Ringheims Auto Member since:
2005-07-23

I may not work in the same field as you, but you're saying that proprietary software is easier to use and better documented, like that is true IN ANY CASE. I have used welldocumented free software, and bad documented and hard-to-use proprietary software. It's not like free software is SUPPOSED to be hard to use and badly documented, in many cases it's rather not, and the contrary is true for many proprietary softwares.

Your employers SHOULD be worried about if what they pay for is Free or not, given that it could go out of business anytime... In the case of Free Software that of course doesn't matter as much. I think there are advanteges of using Free Software beyond the ones you mention.

Not to be getting too much on that "hallelujah-stuff", it's just that I'm very happy to be able to use all of this Free Software, and it's definetely a very big advantage in itself that it is Free.

Reply Score: 2

RE[7]: Lets get real
by Robert Escue on Sun 1st Oct 2006 12:59 UTC in reply to "RE[6]: Lets get real"
Robert Escue Member since:
2005-07-08

There are F/OSS products that are well documented and easy to use (apache, Nagios, OpenSSL, etc.) and we use this software daily. And there is commercial software that their documentation blows goats (Verity). And I have had the joy in trying to figure out how a commercial product works when the documentation doesn't cover it.

As I stated in my reply to SEJeff, we are not just a *nix shop, we also have a lot of Windows servers. So our solutions to particluar problems cannot be just *nix. For example we use Nagios for monitoring because it works with both *nix and Windows machines. We also have a desire to do Asset and Configuration Management for several thousand assets. Can a F/OSS piece of software do this across several platforms, possibly. My recommendation was Tivoli, something that I know works and can be extended to perform a variety of functions besides Asset Management and CM.

We do not have the programming staff to modify and write code for F/OSS software to make it work in our environment, and to write it for multiple platforms. We also have to deal with US Government security requirements, one of those being a code review. Most of the security people I have come across are not programmers so asking them to perform a code review of a particular piece of software is a waste of time. And I have been at the losing end of too many arguments trying to get F/OSS software used for a particular project, so I choose the commercial product.

I had the opportunity to listen to RMS at HOPE 6, and while I agree with him about free software at an individual level, I do not think his ideas translate well in a corporate environment.

Reply Score: 1

RE[6]: Lets get real
by SEJeff on Sun 1st Oct 2006 01:51 UTC in reply to "RE[5]: Lets get real"
SEJeff Member since:
2005-11-05

Robert, after spending 10 years building a configuration management framework, supporting our systems is easier than using the (proprietary) redhat network. Also, we use the RHEL source so (in effect) we are using RHEL customized with various compile flags and some of the bloat removed.

In our situation, paying a vendor doesn't make sense as we can deploy a new production server from blank harddrive to ready for requests in ~15 minutes thanks to homegrown tools. Look at google... They are the second largest linux server builder under dell. It is common knowledge that they use their own homebrew version of Linux and not a COTS solution. Assuming you have the expertise, Linux affords you the flexibility that a proprietary solution does not.

Note that I modded you +1 because I agree you should use the right tool for the job. Sometimes it makes sense to use proprietary software and in ours, it really doesn't.

Reply Score: 3

RE[7]: Lets get real
by Robert Escue on Sun 1st Oct 2006 12:39 UTC in reply to "RE[6]: Lets get real"
Robert Escue Member since:
2005-07-08

Thanks, and I modded you up as well. In your case you have the talent and the time to develop a solution and that's cool. We are a small shop that has to deal with a lot of machines, including Windows servers which means in some cases we cannot just think *nix.

That is one of the reasons why I tend to go commercial first, is because most (but not all) F/OSS solutions tend to be for *nix only and we do not have the Win32/64 programmers to massage code.

Now if only all discussions on here would be like this!

Reply Score: 2

Revenue is an idiotic metric...
by zztaz on Sat 30th Sep 2006 03:11 UTC
zztaz
Member since:
2006-09-16

...for measuring the success of software.

In related news, revenue from the sales of light bulbs continue to outpace sales of sunlight; Sol no threat to Philips.

Reply Score: 5

Narrow View
by elsewhere on Sat 30th Sep 2006 03:12 UTC
elsewhere
Member since:
2005-07-13

Revenue numbers are of interest to Wall Street, but mean nothing to me since they provide no context. Unit shipments would be a little more relevant for truly measuring growth or decline.

Even so, they still wouldn't tell the whole story. Appliances are replacing conventional servers in certain areas, and will continue to do so. Looking at areas like NAS, security or network management you'll find applications that once would have been the domain of Unix, migrated to NT/linux, and now being sold on pre-configured turn-key application-specific devices. In the majority of cases these are simply conventional x86 platforms with conventional hardware based on linux. I doubt Gartner is taking these into account, but if you look at unit sales combined we're talking about millions of appliances being sold worldwide. That number will continue to grow, and in a fair number of cases represents the displacement of what would otherwise have been a conventional server sale.

There can be a surprising number of linux installations in the average data center even if none of it is running on a conventional HP or Dell server.

As is usually the case with Gartner reports, numbers don't really tell the whole story. And trying to apply conventional metrics to linux doesn't always work due to it's somewhat unique flexibility and accessibility.

Does this mean I think linux is going to drive Windows or Unix into the ground? No. I simply think the makeup and infrastructure of the average datacenter is going to continue to evolve (big surprise) and using projected server license revenue as an sort of an indicator is a little too simplistic.

Anyways, just my interpretive 2c.

Reply Score: 3

blah
by theGrump on Sat 30th Sep 2006 03:46 UTC
theGrump
Member since:
2005-11-11

first of all, who the hell would waste their time attending the "gartner open source summit"????

second, it must be clear by now that the people using and developing linux do not care if they are outnumbered 3 to 1 or 100 to 1. why do people keep brining this up? ford dirvers outnumber maclaren drivers - what losers they must be.

Browser: ELinks/0.11.1-1-debian (textmode; Linux 2.6.17-2-686 i686; 90x32-3)

Reply Score: 5

the key word is revenue.
by graigsmith on Sat 30th Sep 2006 06:25 UTC
graigsmith
Member since:
2006-04-05

the key word, is revenue. so mabey linux companies won't have more income. but im sure more people will be using linux lol.

Reply Score: 1

RE: the key word is revenue.
by zztaz on Sat 30th Sep 2006 07:24 UTC in reply to "the key word is revenue."
zztaz Member since:
2006-09-16

More exactly, sales revenue. Which is not the important number.

Most software is not developed to sell to other people. It's developed for internal use, or embedded within some other product. As impressive as Microsoft's sales numbers are, the market they dominate is actually one of the smallest parts of the overall software economy.

And that's just the acquisition cost, when we all know that the support cost is usually higher over the full lifecycle.

The revenue number that is important is the revenue generated from the use of software, not the sales of software. The focus needs to be on users, not providers. All software is paid for by the users, whether they built, bought, or traded for it. Looking at only sales numbers ignores the larger build or buy question. It ignores support. It ignores TCO.

George Weiss is looking for his car keys under the street lamp because the light is better there. That's not where he lost them. Sales numbers are easy to find, usage numbers are not. Finding out what software is being used requires actual research, unfamiliar territory for Gartner.

If companies can make more money using Linux than using Windows, that's what counts. They will pay for Linux, but in ways that are difficult for Gartner to track. Sales revenue does not tell the complete story of Apache, or Java, or sendmail. The list goes on and on.

By sales revenues, Internet Explorer and Firefox are no threat to Opera. By sales revenues, web designers should ignore IE, which has effectively zero sales revenue.

Reply Score: 5

Sales != Deployment
by makkus on Sat 30th Sep 2006 06:57 UTC
makkus
Member since:
2006-01-11

We have two Red Hat servers running critical systems, besides that we have 10 Workstations and 6 servers running Fedora. I think linux sales are becoming flat because of more in-house experience will lead to non-licenced based linux systems. Why should I licence my linuxes, I never call them. The two Red Hats are because the MRI manufacturer (Bruker) demands it for its hardware. the 16 Fedoras will never get counted in any sales and I now a lot more places in our Academic hospital where they run Fedora or Ubuntu

Edited 2006-09-30 06:58

Reply Score: 4

Point is competition
by vegai on Sat 30th Sep 2006 07:30 UTC
vegai
Member since:
2005-12-25

Beating Microsoft is not the point (except, apparently, for the Ubuntu folks).

Point is that the industry should be a place where healthy competition can take place. I think we're getting there.

Reply Score: 3

Hidden market share
by moleskine on Sat 30th Sep 2006 10:30 UTC
moleskine
Member since:
2005-11-05

The fact that Linux exists does damage to Windows.

Linux provides countries all over the world with an alternative to Windows they can develop according to their own needs. I believe there is even a Bhutanese version of Linux now, based off Debian. The same is true for ngos and non-commercial outfits of all stripes who may have ethical or financial reasons for not wishing to go Windows.

Linux is a superb platform - even if only viewed as a marketing platform - for open source products. Each one of these has taken business off Microsoft, whether server software like Apache or Samba or desktop software like Open Office or Firefox.

Linux is a showcase for the open source development model. Not the least benefit of this is that it has put the spotlight very firmly on Microsoft's attitude to software patching and security.

In other words, Linux has had a profound impact that goes way beyond its market share. And its market share is probably consistently underestimated because of the "no-registration" download and install model.

Decent gui administration tools might well do a lot to further Linux's popularity on the server, but that's another story.

Reply Score: 4

They must LIKE Microsoft....
by Ringheims Auto on Sat 30th Sep 2006 11:41 UTC
Ringheims Auto
Member since:
2005-07-23

Dunno if MS is paying them or not, but in many cases it DOES seem that Gartner actually LIKES microsoft, for one or another reason, and would like to see any competition vanished. They never mention Apple, and talk about GNU/Linux almost as a threat to ordinary users who they think very much prefers windows (but who don't have a clue what they're using/prefering anyhow....).

Gartner, get a clue, and don't expose your microsoft fan'ism that much....

Reply Score: 1

walterbyrd
Member since:
2005-12-31

Unknow to the bozos who do these kinds of studies, not all Linux is redhat, or even novell.

A lot of people use debian, or slackware, for their servers. In fact, a lot of people use linux because it's free.

One of the great attractions of linux is that it's much less expensive than proprietary alternatives. That means that revenue is poor measure of market penetration. Whoever commissioned this "study" probably knows that.

Reply Score: 1

Wow, I wish I could see the future too
by deficite on Sat 30th Sep 2006 13:44 UTC
deficite
Member since:
2006-09-30

What ticks me off about crap like this is that people think that statistics and stupid biased opinions can tell the future. A lot of things can happen in even one year that can completely change the world, and some guy getting paid to say his opinions is not going to be able to predict it. People say Nostradamus predicted 9/11 and so on and so on, but despite that (whether it be true or not, which I doubt) nobody would predict that the aftermath and effects of that attack would come to be. Statistics will never be able to tell the future in that way.

Has anybody noticed how many of these things are flooding the internet lately? It's almost like it's a conspiracy or something. It's ridiculous how many people are writing stupid articles like this saying that "I know the future, I have seen it. STFU Linux users, because you're all hopeless and retarded. Switch to Windows or get caught in the wake! MUHAHAHAHAHA!" and then getting paid for it. Maybe I should write a paper with a bunch of phony numbers used out of context, etc. and get paid for it?

Reply Score: 1

Deficent windows model at work again...
by buff on Sat 30th Sep 2006 14:06 UTC
buff
Member since:
2005-11-12

Articles like this always make me chuckle a little bit. It is always the same thing: Linux will never win since it isn't Windows-like enough. Real linux users that know what is going on don't care about this. I actually like the state of desktop linux the way it is. A real article would address some of the major points that prevent universal adoption of Linux by the public:
1. Microsoft dominated workplaces, software distribution in retail stores.
2. Higher technical knowledge needed to setup the Linux configuration. Knowledge of using a shell is still needed
3. Lack of real transparency with working with Office files right out of the box. For example Macros in Word fail to work right in Open Office preventing my company from using OO for certain tasks.
4. Cross-platform gaming is getting there but not quite there yet shying away Windows users.

Reply Score: 2

Gartner, Gartner, Gartner...
by 0xbadbeef on Sat 30th Sep 2006 10:44 UTC
0xbadbeef
Member since:
2005-11-12

Gartner predicts that Linux will not much of a dent in Window... Hmmm, isn't Gartner the same bunch of wankers that mispredicted the success of Itanic by 95% and promissed that Itanium will take over the world by 2003? Gartner would be the last place I would listen for any sort of projections. I on the other would go as far as saying that it is inevitable that Linux *will* make a dent in Windows marketshare especially in the developing world.

Reply Score: 2

RE: Gartner, Gartner, Gartner...
by raver31 on Sat 30th Sep 2006 23:22 UTC in reply to "Gartner, Gartner, Gartner..."
raver31 Member since:
2005-07-06

Yep, I will believe your prediction over Gartners.

Escpecially in a few years time when the recipients of the OLPC start releasing software back to the community.

Reply Score: 1

Skilled labour
by shiro on Sat 30th Sep 2006 11:22 UTC
shiro
Member since:
2005-07-20

Ah.. the old Windows vs Linux at server market again.

This issue isn't about Linux is better than Windows, it's more about decision making from upper level guys; guys who usually say something like:

"You know, Linux is good and sexy, but let's stick to what we usually have and just upgrade to its newer version rather than re-do our infrastructure. Besides our internal web architecture is based on ASP."

Migrating is one of many problems. .NET seems to be a good (and valid) choice for enterprise too why bother send engineers on PHP training (or PHP conf or whatever). That cost more money and time to deliver a new infrastructure.

The first decision of how a company infrastructure will be built will stay there forever in most cases.

Migration is not just about cost, it's about time and result. Will the result be better? How can you make sure we can migrate and all. It's not like we're migrating from Windows NT or Windows 2000 with ASP to Windows 2003 running ASP.NET.

Skilled labour is the other reason why using Linux might be a problem. It's harder to find a good Linux kernel engineer (in my opinion, a good kernel engineer most likely is the better person to be a sys-admin or net-admin).

Google runs thousand of Linux servers with the backing of their superior engineers, not just a bunch of guys who took Linux books and study its commands and shell.

Reply Score: 1

So...
by Sphinx on Sat 30th Sep 2006 17:50 UTC
Sphinx
Member since:
2005-07-09

How come nobody measures the Microsoft threat to Linux?

Reply Score: 1

RE: So...
by Babi Asu on Sat 30th Sep 2006 18:36 UTC in reply to "So..."
Babi Asu Member since:
2006-02-11

Because it is silly.

Reply Score: 0

RE[2]: So...
by Sphinx on Sun 1st Oct 2006 06:01 UTC in reply to "RE: So..."
Sphinx Member since:
2005-07-09

I can only assume you've never tried to compete against Microsoft.

Reply Score: 1

RE: So...
by raver31 on Sat 30th Sep 2006 23:27 UTC in reply to "So..."
raver31 Member since:
2005-07-06

Simply because a company, (Microsoft), cannot threaten an ideal, (Linux)

Linux has no one target for Microsoft to go after, it is like the hydra. Kill one head and another sprouts up in its place.

If by some strange miracle, Microsoft makes all companies and individuals to stop making Linux, some day in the future, a spotty guy in his mums basement picks up the source code, and starts it again. Linux is in the open, and it can never be put back.

Now, if Microsoft was to go under, Windows would die. No-one could bring that back.

Reply Score: 2

RE[2]: So...
by happycamper on Sat 30th Sep 2006 23:45 UTC in reply to "RE: So..."
happycamper Member since:
2006-01-01

Now, if Microsoft was to go under, Windows would die. No-one could bring that back.


any company can buy the complete rights to windows souce code, and name, and can start offering windows.

Edited 2006-09-30 23:46

Reply Score: 0

Let me rephrase the headline:
by Temcat on Sat 30th Sep 2006 13:45 UTC
Temcat
Member since:
2005-10-18

Linux Is Not About Doing Damage to Windows

That explains all, doesn't it? ;-)

Reply Score: 3

I used to care, but...
by h3rman on Sat 30th Sep 2006 19:55 UTC
h3rman
Member since:
2006-08-09

I have to admit, I used to care about whether Linux could see growth at - preferably - the cost of Microsoft's (in)famous operating system.

But then you realise, all these figures, articles, research papers, battles, discussions, predictions of what's to come.., all that is only because of the fact that people are so fond of discussing and watching, and partaking in, conflicts - real or imaginary.
Isn't it great, a cosmic battle between two empires, one very visible, corporate, monopolistic, another less visible, independent, more liquid and freaky?

But that has nothing to do with what it's all about.
It's all about quality. For quality to eventually prevail, not much is needed. Just dedication, and patience.

Reply Score: 2

the actual facts please
by happycamper on Sat 30th Sep 2006 23:36 UTC
happycamper
Member since:
2006-01-01

Microsoft Windows will not suffer irreparable damage on the server side at the hands of the Linux operating system.

I think this sounds more actual.
for sure Microsoft Windows will suffer irreparable damage on the server side at the hands of virus, trojans,spyware,etc.

Reply Score: 3

RE: the actual facts please
by raver31 on Sat 30th Sep 2006 23:49 UTC in reply to "the actual facts please"
raver31 Member since:
2005-07-06

hahaha that one was very good. I will give that a +1 before the fanboys mod you down

Reply Score: 1