Linux Growth In The Workplace Slowing

BrainSurgeon writes "According to a Business Week article Linux growth numbers have slowed for the first time since SG Cowen & Co. began tracking it on their survey. The biggest reason for the slow down according is due to the hidden cost of consultants." From the article: "That doesn't mean overall Linux use is slowing. The survey only shows that a smaller number of companies not using Linux plan to try the software than in previous surveys. Most analysts expect Linux use to grow at the companies that have already rolled it out -- and do so at a healthy rate. And analysts say Linux is picking up steam outside North America, which the Cowen survey doesn't cover."

How does it compare with Windows growth?

[bc90021]

Is it slowing too? In which case, this is a sign of overall growth slowing... or is it just a Linux thing?

I am not surprised

[Skiron]

BACK TO MICROSOFT. Take Independence Air, a low-cost Washington (D.C.) carrier that had been running the reservation system on its Web site with Linux. The company, which uses Microsoft's Windows operating systems in most other pieces of its business, needed to hire consultants who could write code for Linux, since its Windows developers couldn't.

What can 'Windows developers' do? Use a mouse?

And if this statement is to do with the code running on a web server (Apache, I presume), then even more so I feel they hired the wrong 'developers' to begin with.

Just more FUD - move along.

Liars, Damned Liars and Statisticians...

[the_rajah]

in that order. You can make the numbers say anything with sufficient qualifications. Oh, and it's only in the US, not world-wide. Read the article carefully. It doesn't mean Linux adoption on a per user basis is slowing, as I read it, the growth rate of new companies using it is slowing. Hardly time to get out the black armbands..

Growth Slowing but still growing

[Mutilated1]

Well just because the growth is slowing doesn't mean that its not still growing. The HPs and IBMs of the world already are using Linux, but the smaller businesses aren't really leading the way with Linux. At my company ( about 500 employess ) we are just this year using Linux for some servers, and I know several other companys that are just starting to use Linux too. So even though these companys are small, as machines age more and more of them will be replaced, and more and more of them are being replaced with something other than Microsoft. Maybe it won't happen overnight, but it will happen none the less.

Re:I am not surprised

[Anonymous Coward]

I can't read a text file!!! I can't use an ODBC interface to our server backend!!! This is way over our heads!

I don't do a lot of programming anymore, but integrating an opensource backend with a Windows frontend is a cakewalk compared to going the other way.

I'm guessing that their programmers do nothing more than generate some stupid crystal reports from a SQL database.

Standard Gartner hype curve

[jurt1235]

It can just mean that linux is normalizing. People are not able to hype it anymore to get it into a company. This can be the best thing, since where it gets in, it will stay, and that way slowly gain market share.

Multiplatform Deployments

[maynard]

Take Independence Air, a low-cost Washington (D.C.) carrier that had been running the reservation system on its Web site with Linux. The company, which uses Microsoft's Windows operating systems in most other pieces of its business, needed to hire consultants who could write code for Linux, since its Windows developers couldn't.

"That cost was killing me," says Stephen Shaffer, Independence's director of software systems. After eight months, he replaced the system with Windows and a batch of other Microsoft applications, which he believes will cut his costs by 70% a year.

Naturally, Microsoft sees the Cowen survey as proof that Linux is finding resistance. "This data completely validates what I've seen," says Martin Taylor, Microsoft's general manager for platform strategy. Not only is Linux maxing out on Unix users but it's not finding new customers among stalwart Windows users, he says.

These statements are skewed to show that Independence Air's Linux deployment cost too much in consultant fees, and therefore Linux is "expensive" to deploy in comparison to Windows. But they really say no such thing. Independence Air's problem was not its Linux deployment, but the fact that it chose to deploy a small part of their infrastructure without in house knowledge. They already had hired a Windows skill base, and therefore the comparison in utility between their Windows skillset for the entire Windows deployment against a small Linux deployment was bound to come out poorly for Linux. One sees savings with Linux in scale, not individually. Deploy hundreds of hosts and you'll save huge. Deploy a few hosts to drive a small piece of corporate infrastructure and not only will the savings be marginal, but you may have to hire external help to support the deployment.

So. Don't deploy Linux for small tasks if you're already heavily invested in an alternate technology. Duh. But to claim poor savings across the board as a result of this anecdote is simply stupid. With in house Linux (or UNIX) personnel and a large deployment - of course you'll save big. Which is why the UNIX houses have dumped commercial UNIX desktops for Linux. And why so many have dumped all their small UNIX servers for Linux (and BSD) on Intel. Because it's cheap. Very cheap (and cost effective). --M

Read between the lines

[coolsva]

Only 7% of outfits with no Linux servers plan to add some over the next year.
Can we safely assume that we are approaching the limit of diminishing returns, all those who are amicable to convert, have already done so.
Also, this doesnt mean Linux is slowing/stopping. Companies with some servers would definitely go forward with more, thus growing the overall Linux implementations.

Simple

[kf6auf]

It's really quite simple: the first derivative of linux use (growth) is positive, but the second derivative (acceleration) is negative. Let's just all hope that the third derivative is positive.

another Microsoft sponsored study

[Anonymous Coward]

This time we are looking at a more indirectly sponsored study by Microsoft. You might want to look up who is behind SG Cowen & Co. LLC and you'll see they are largly owned by Société Générale, a large French bank that is one of the major investors of Microsoft.

Sorry, but as everyone knows Linux is gaining market share very rapidly. Nice try...

It makes sense...

[freality]

The *required* cost of using Linux in the workplace is, of course, $0. Just hire people who've been using it for years and you don't have this big scary learning curve. They, like any other long-term OS user, knows how to manage the version stream and keep focused at the same time. Except with Linux, there's no other costs. It's free.

The way it actually gets to cost something is driven by a market. The question isn't how much free Linux costs. The question is how much Windoze-imitation Linux costs. Companies who are accustomed to shelling out big bucks for Windoze will shell out just a bit less for anything else that does the job and call it a win.

So of course, for many things, Linux does the job.

And then there's RedHat to charge just a bit less.

That's all it takes to bring Linux TCO up to Windows range. I've seen it happen, with my very own eyes. I've even seen a company pay *more* for Linux than Windoze.. *and be happy with it* because Linux is higher performance for many server applications.

"Unbelievable!" I thought. But it's the market and the expectations that set it up.

No matter that you can d/l and install Fedora to do just the same job in less the time than it takes to call a RH consultant to get even a quote. You just shout "Risk! Risk! Risk!" enough and you get your IT department a fat budget and get to wear a Linux T-Shirt.

It's like saying A bird in hand is better than two in the bush. "Sure we could all become Linux experts, but maybe we'd fail!"

Businesses understand and practice outsourcing intelligence all the time. That's their bird in hand.

Re:I am not surprised

[bluGill]

Many windows developers have never considered the idea that all the world isn't the latest version of Windows running on a 386. (That pointers are the same size as integers, and you have little endian machines) They might have gone for SSE or some such 3d extention, but only if they need it.

Unix developers tend to target more machines. This results in a lot of little assumptions they get right. They are likely to have played with several different windowing APIs. (Often starting with Motif, but certainly qt, GTK, and WxWidgets) This results in a code style that is flexable in the face of different machines. It is also more expensive to develop simple software like this. (though because it is well designed,complex software that can't crash is cheaper. Most people don't need this level from their applications)

Windows developers tend start from VB, which in itself ruins any chance of ever becoming a good programmer because it encourages bad habits. Even when they are good, they have not had to deal with the issues that come up when your machines are not all the same family of CPU, with the same API, so they don't make design choices that allow for those differences. This results in most of their code not being as stable (though there is no obvious connection between the two).

Thus Windows developers have more trouble crossing over to Unix, than Unix developers have crossing to Windows. Note however that either direction is painful and difficult. Add in that many Windows developers are not any good to begin with (though there are many bad Unix developers as well), and it is easy to see why there would be problems.

Answer: No, The Penguin is Fine

[ThisIsFred]

It's MS marketing in action, no more. Apparently they're clearing the path for some new announcement, so the shill army has been deployed.

Selected excerpts:

But a new report suggests that growth rate could be about to slow.

"A" report "suggests" it's about to slow. Windows growth is zero, but Linux growth in 500 corporation buyers, limited to those who a) plan on buying new servers, and b) plan on having Linux installed, may slow to 7% growth.

The survey only shows that a smaller number of companies not using Linux plan to try the software than in previous surveys.

Have all the companies that began adopting stopped or reversed adoption entirely?

Most analysts expect Linux use to grow at the companies that have already rolled it out -- and do so at a healthy rate.

But not these guys, apparently.

Brosseau thinks Linux won't find as many new customers as in the past in large part because it has already tapped the market that includes its most likely buyers: Unix users.

So this guy says. If anything, it shows where those marketing Linux-based solutions aren't reaching out to customers with other needs. Linux does a fine job as a low- to mid-range file and print server, document-server and company intranet web-app server. That's about as easy a drop-in replacement that you can get. However, thinking is required.

What's more, some companies that experimented with Linux hoping for big cost savings found that the open-source software wasn't what they expected. "Some of the bloom is coming off," says Brosseau.

Who are these companies, and what do they do? How did they implement it? Where is the report with cited sources to back up this statement?

Take Independence Air, a low-cost Washington (D.C.) carrier that... needed to hire consultants who could write code for Linux, since its Windows developers couldn't.

...

...replaced the system with Windows and a batch of other Microsoft applications, which he believes will cut his costs by 70% a year...

Ouch! This seriously undercuts the credibility of MS-certification, since this crew doesn't seem to know how to read. Of course there will be greater costs during the switch, but will that remain at 70% above the current costs forever? Also, this is an anti-argument for Microsoft, because this same company is obviously going to have a hard time dealing with the API changes coming in Longhorn. If anything, this message is: "It will cost a lot more to move away from a Microsoft solution if you buy into it, and your applications won't be portable to other platforms. Microsoft can't grow with your business, and you'll be stuck with them!"

Naturally, Microsoft sees the Cowen survey as proof that Linux is finding resistance. "This data completely validates what I've seen," says Martin Taylor, Microsoft's general manager for platform strategy. Not only is Linux maxing out on Unix users but it's not finding new customers among stalwart Windows users, he says.

This guy works for MS marketing strategy. Did we really need to quote him? Also, his statments contradict the concessions made by the author of the article. Who is right?

Siemens, for example, has Linux servers handling some of its firewall and communication applications. Now, it's pushing Linux into its data center, heavy-duty computing that hasn't been the open-source operating system's forte.

It hasn't? Linux has been the platform on which to build a "super"-computers with commodity hardware. It's also a popular platform for serving massive loads in heavily-used websites. It can do more. The author also fails to point out that Microsoft, despite its major marketing effort to get Windows into the "data center", is encountering resistance. Let me put this forth: If operating systems such as Solaris and AIX are in the "data center", and Linux is replacing UNIX, where is Linux going?

Re:The cult of UNIX strikes again

[Dollyknot]

Imagine going through life with not being able to talk to people in spoken/writen language, where all you could do is click your fingers and point at something, anagalous to going to a foreign country an not being able to spick a da'lingo. Where all you can do is vaugely wave your hands around, point at things and hope they get your meaning, The truth of it is, the CLI is much more akin to speaking to your machine instead of making threatening gestures at it via a GUI.

I'm tempted to say 'point and click' my arse, but might it might give you the wrong idea :)

Yes on the surface in some areas Linux might look a bit like a GUI spatchcock, but its heart is golden, engendered by thirty years of evolution and stress testing, windoze has nowhere near the pedigree.

We live in a black box society where people think all that is necessary, is to have the correct label for everything and to know which button to press. There has the inherent danger in this process, of dumbing everyone down.

Truth to tell, it is not about UNIX Linux or Micro$oft, it is about open source versus closed source, this has nothing to do with GUI/CLI.

Read the man [totse.com] (Richard Stallman)