Slashdot Mirror
Amazon: Linux Saved Us Millions
Ian_Bailey writes: "ZDNet news presents another chapter in the Windows vs. Linux debate. Amazon.com claims that by switching to Linux, they were able to "cut technology expenses by about 25 percent, from $71 million to $54 million."" Lots of little bits in there. Nothing really new, but it's still nice.
isnt this more about Unix vs Linux than Windoze vs Linux. Hey, dont let the facts get in the way of a good headline!?
Funny, that sounds very similar to what RedHat does as well.
--It's Pimptastic!--
How many patches and updates have Microsoft published in the last year? And how many of those were pulled, because they weren't tested properly? I haven't had to re-install Linux on any of the boxes I admin since we went to RH 6.2. The MSCE-in-training down the hall can't say that about the last three months on his Windoze boxes. Imagine doing that for 100, 1,000, or even more! What fun! Thank you, Microsoft!
In a corporate Linux environment you still have to pay for the staff, sysadmins, distributions [most likely Amazon bought a few copies of what ever linux.dist they have], etc..
Have you ever noticed the prices of Windows NT Server and their various other BackOffice applications? You're talking many hundreds of dollars per server for software alone. Plus you need client licences... This might just be a drop in the bucket for some businesses but when you have lots of computers, the costs add up.
I'm all for Linux too, but lets not use any excuse to promote it. The OS is still very premature and doesn't really compete with Windows in terms of END USER usability. Sure Linux may be a faster/stabler[?? VM problems?] OS, but it is not an easier to use OS, and in the end, for 99% of the users that is what counts.
What end users? The only ones that would probablly be able to tell that the computer was running Linux are the system administrators. And for ease of use... administering Windows NT isn't an easy task.
Will this battle ever end? 5 Years ago I heard people telling me that Linux was going to take over as a desktop operating system, and squash Windows 95 technology. Well, I didn't believe them at the time, and I still dont see it happening. I don't think we should really worry about making 1 operating system the most popular. Just make it work, and make it interoperable with other operating systems. I like to use linux because it gives me alot of versatility, and also gives me a better view of my network when something goes wrong
On the other hand, Microsoft's Media Player is pretty pimp considering it will open about any video compression scheme I throw at it.
The operating system you use is your choice. Let the less savy use Windows, because thats what they want, an easy OS.
Can all fish swim?
As a corporate entity, sure, but it's been my experience that a switch from M$ to GNU/Linux requires an internal evangelist. The guys in charge of running the machines have to want to do it, otherwise they can come up with any number of reasons why it's not viable right now.
Somebody in there is GNU/Linux friendly. Let's raise a glass to their health.
If you were blocking sigs, you wouldn't have to read this.
in the Wall Street Journal? Maybe we could all chip in for something like this:
[a picture of federal marshalls carting computers away from from a business, horrified managers in the background]
Complicated licensing and expensive audits could land you in legal hot water and cost you your business. Linux will save you money and give you peace of mind. [Add examples of companies such as Amazon that have moved to Linux.]
Umm, yeah cnet put in some stuff that MS has SAID IN THE PAST, but this paragraph should explain what the real price cut was from.
HP has been working with Amazon since October 1999, Balma said, but the big contract win came in May 2000, when HP announced its [linux] systems would replace Unix servers from Sun Microsystems.
They replaced Solaris boxen with Linux boxen. This, actually, has nothing to do with Microsoft.
CNet just put it in there to hype the article.
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
However, when you get down to it, it is about money. They don't really give a shit about Linux itself. They don't have feelings for it. Don't forget that. It is about the money.
Well, yes...?
[Insert any company name here] is actually more concerned about finding an effective, sustainable way to reduce their operating costs than about the social, spiritual, and aesthetic qualities of they tools that they employ to achieve that reduction.
I may be interpreting webwords' comment incorrectly, but it sounds to me like "don't trust those corporate drones, they're only using Linux 'cause it's practical...they *should* be committed to Linux regardless of any financial considerations."
If a major corporation makes the decision to switch to Linux because it is the cheapest and most effective solution for them, it means that those people who do "have feelings for" that operating system are doing their job...making the operating system an justifiable option for the many, many people who just want an OS that allows them to run their business effectively.
* * *
It is a dada story -- it has no moral.
But there are hidden costs to Linux, Microsoft argues. ....
With Linux, customers "end up being in the operating systems business," managing software updates and security patches while making sure the multitude of software packages don't conflict with each other," Miller said. "That's the job of a software vendor like Microsoft."
Yeah, like Microsoft does a good job of that. Like how many times do I see "This program has performed an illegal operation and will be shut down"?
The reduction was attributed primarily to Amazon's "migration to a Linux-based technology platform that utilizes a less-costly technology infrastructure, as well as general price reductions for data and telecommunication services due to market overcapacity," according to the filing.
My concern here is in finding out how much of that savings can be attributed to the switch to Linux and how much can be attributed to reduction in data and telecomm services? Real data here would be interesting, but it's probably just not available.
Thirdly, in many cases companies don't have to pay extra licensing fees for the computers that connect to Linux servers. And finally, Linux is often used on inexpensive Intel computers, sometimes generic "white box" machines and sometimes older computers seeing a second life.
This is a real hidden gotcha, and in many cases a tremendous potential for cost savings. If only companies would truly look at this item before investing. Servers with "per seat" licensing can really escalate costs, especially for a business like Amazon. You think that by buying one program, your costs are over and you're done with it. But as your customer base goes up, you have to start paying additional licensing fees. Budgeting for this kind of stuff is difficult at best. Microsoft is certainly guilty of this, but they are joined by other powerhouses such as Oracle. To their credit, many of these large companies offer some sort of "enterprise" deal, but it usually has a whopping price tag associated with it. In my opinion, it's much more fair to sell by the server. If I want to try to cram 2000 users onto a single server, it's my responsibility to deal with the resulting problems.
But there are hidden costs to Linux, Microsoft argues. "I think a lot of customers are lured by the apparent low price of Linux," said Doug Miller, director of competitive strategy for Microsoft's Windows division. "They don't have a real issue with Linux, but it ends up costing them in the long run."
With Linux, customers "end up being in the operating systems business," managing software updates and security patches while making sure the multitude of software packages don't conflict with each other," Miller said. "That's the job of a software vendor like Microsoft."
Oh yeah, I've written soooo much Operating system code since I started using Linux. The last update I did (over a year ago), I didn't even feel a single urge to recompile the kernel. And what's this about managing software updates and security packages? Exactly WHAT does Microsoft do for me that I don't get with RedHat's up2date or Debian's apt-get? Software packages conflicting with each other? What does Microsoft offer to take care of this problem? I've certainly had it enough times in the past with Windows software that I could have used some help. Boy, those two paragraphs are the biggest bunch of baloney I've seen in a while -- and I was in Germany for six months!
While Red Hat offers some of those services, it's difficult to ensure that software packages updated frequently by hundreds of people around the globe work well together, Miller said.
Really? I don't think I've EVER downloaded a single package from RedHat that didn't work just fine with all of the other install packages from RedHat. Anybody else had any problems with that? I guarantee you that RedHat does at LEAST as much testing as Microsoft. Let me remind you of NT 4.0 SP 6....
Among those forces: the coming version 6 of Sun Microsystems' StarOffice package of office software, which many believe will be a more capable product than the bulky current version and thus a more credible alternative to Microsoft's Office; burdensome Microsoft licensing fees during a time of economic austerity; and the overall price tag of Windows and Office.
OK, I'm not sure that I can agree that StarOffice is or will be more capable than MS Office, but with the current economic times, the price is certainly much more attractive. And if you look at what most people actually use an Office Suite for, you'll find that almost all of them will more than have their needs met with Star Office 6.0.
The study concluded that Linux applications could provide solid alternatives to nearly every Windows application, with the possible exception of the scheduling and e-mail integration of Microsoft Outlook.
And a nice WYSIWYG, comprehensive web-design suite like, say, Dreamweaver, would be a nice addition to Linux. Anybody try out IBM's WebSphere Home Page Builder for Linux yet? Scheduling and e-mail integration is one thing I wish OpenOffice (OS version of StarOffice) hadn't dropped from their focus. Even though there are some nice e-mail and scheduling programs, it would be nice to have tighter integration with my other office software.
"Staying in compliance with licenses is something a lot of companies are scared of right now. It's more difficult, and the ramifications of being out of compliance are becoming more and more onerous," Robinson said. "As of the last year or so, Microsoft has been going after companies where they've gotten tip-offs or had other suspicions."
This is another big one. I heard a radio advertisement this morning offering to help companies get in compliance during the grace period. They through out all those scary numbers like $150,000 per violation. You absolutely know you've got a problem when agencies can actually derive their entire revenue base from helping people manage the complicated licensing issues that Microsoft has created. This whole thing is exactly what prompted me to switch to Star Office on ALL of my computers. I had licenses for the versions of MS Office I was using, but I didn't know what scheme they would think of next.
"We are a commercial software vendor. That's how we earn revenue," Miller responded. "Our goal is to be properly compensated by customers for our software."
And to make additional money off of existing customers by "clarifying" the terms of the license to them and forcing them to upgrade and pay additional licensing costs when they don't want to....
GreyPoopon
--
Why is it I can write insightful comments but can't come up with a clever signature?
They don't tell us.
... well that's another matter entirely.
Hell this is INTEL.
The most likely answer is that they're cutting over their design stations from Solaris or HP-UX.
If there's tons of money to be saved it's on the IT support for Engineering. This is where the truely ghastly sums of cash get spent; and it's where Linux is starting to pound on Solaris and HP-UX.
It's also much easier to sell the idea of a Unix-to-Unix conversion to the managers -- engineers tend to love this sort of thing. Where-as selling and NT-to-Unix conversion
With them, you pay throughout the short-term, the forseeable future, and even the long run.
.NET, if you bought an OS from them in 98, then you needed another one in 99 (98se), another one in 2000 (Me), and another one in 2001 (XP) plus numerous hardware upgrades. So, the average cost of those software upgrades is $75 Per Boxen, and probably $100 in hardware upgrades, again, per box.
You're modded up as funny, but you've hit on a key point... from the article: "But there are hidden costs to Linux, Microsoft argues. 'I think a lot of customers are lured by the apparent low price of Linux,'"
I'm not lured to linux by the apparent low price of anything. I'm lured by the actual free price. This whole thing with "you'll pay in the long run" is rediculous. Yeah, you'll pay in the long run. You'll pay to keep your distro up to date - a tech support guy has to type apt-get update every coupla days. Or, heaven forbid, once every 2 years, you'll want to bust down $60 at best buy for the "enterprise" edition of whatever distro you want.
Now, microsoft... Even before
I just fail to see where the "pay in the log run" comes into linux that it doesn't come into windows.
sig?
I saw a couple of comments, even from Rob, that seemed to say, "We've seen this before" in regards the Amazon announcement. I would like to submit that Amazon's announcement matters, not because of their company size, but because of how they behave. They are not early adopters or innovators, they are a technology risk averse company that bets their business on technology. In short, Amazon illustrates the critical tension facing both the Linux community and Microsoft.
There has never been a technology company to last for more than 20 years on a single family of technologies, and, more to the point, the failure of technology companies has never come from having their dominance in what they do well attacked. Technology companies fail because someone else steals their avenues of growth.
If you look at IBM, it went through waves of changes, starting in the digital age with mainframes, which dominated the marketplace from 1960-1980; selling to enterprise customers digital computers that would dramatically change their business. It saturated the enterprise with mainframes by 1980 and had, starting in the 70's, tried to maintain their growth rate by selling mainframes to middle market (500-5000 employee) companies who had not purchased mainframes.
Along came Digital Equipment Corporation, with the VAX, which just completely took that midrange market by storm, sapping the growth from IBM. IBM built the PC and launched a new market targeted at small business, but Apple, Compaq and a host of clones sought that market and, in the past 15 years, largely took that growth away from IBM.
IBM has been growing it's services business and it is paying off, driving an increasing portion of revenue. They are in year 8 of fantastic growth, but already, they are making noise about trying to sell services to businesses in the middle market; a sure sign that something else is about to come along to meet that need.
Why the history lesson? Because it illustrates the fundamental forces at work that are affecting the Linux and the Microsoft worlds.
The technology industry is characterized by several constraining forces; the innovation force, that seeks the best solution for a given problem, and leverage, the drive to extend technologies from one market to another to extract the best return on investment for that innovation.
Best solution is a subjective term, but in this case, it refers to the solution that is most applicable to a given problem, with the required supportive ecosystem around it and with the lowest cost of aquisition and the cost of ownership over the life of the technology. Hold onto those four points, they will become important.
Microsoft truly came up with the best solution for desktop productivity. Windows was a unique technology in that it brought the ease of use of the macintosh (meeting the test of applicability) that had the lowest cost of acquisition (OEM pricing included it with the computer), the required ecosystem (cheap PC's, compared to expensive proprietary Apples) and a decent cost of ownership (compared to the alternatives at the time, like DOS, which required extensive training).
Fast forward to today. Microsoft is now limited by the slowing growth rate of the personal computer industry, so it seeks to adapt its technology to other markets, in the name of leverage (internally) or compatibility (externally). So we see Windows in the Pocket PC format, where it is touted as an embedded system for extending the productivity brought by your PC. This embedded systems market is large, and fractious, as it extends from cell phones to pda's to robotic industrial arms to game consoles.
Linux is a contender for this market, using our criteria of best. Linux has the best applicability, as it is a modular OS that is compiled for the specific use. Want to use it in a robotic arm? Ditch the graphics processor and X-windows, strip it down to just what you need. Cell phone? Take out large portions of the OS that support complex sound and graphics, devices, hard drives, etc. Game console? Build up the graphics processor support and sound, device drivers and ethernet, get rid of the general use stuff that isn't needed for running really fast games.
Windows isn't nearly as modular, you can turn off functions, but it causes the OS to behave in funny ways because it was never meant to have these things turned off. So, Linux wins the applicability aspect of it.
As far as supportive ecosystem, this is where the battle really lies for embedded systems. Microsoft has brought it's armada of partners to the Pocket PC, to the XBox and to other embedded system projects, but these partners suffer from the same applicability problems that Microsoft faces. Do you really need MS Money running on your PDA, or would a simpler checkbook program that can interface with MS Money easily be better? Do you really need MS Access running on the PDA, or could a simpler program do the trick more efficiently.
In general, it is always more advantageous for the customer and more costly to the provider to innovate for a specific use than to stretch innovations across uses. As the embedded systems market grows, the viability of applications in this space will grow along with it, especially as standards for hardware coalesce.
Between Windows and Linux, the ecosystem criteria is a tie for now, but what about cost?
For manufacturers of hand held devices and specialty use devices, like game consoles, cost is a primary concern. When you are building super computers, the cost per component is a moot point, but for consumer goods, it becomes paramount. Cost of aquisition for Linux is not, as commonly percieved, zero - there is a cost in modifying the OS to get what you need and the cost of support, which is the very business model of Red Hat, but it is substantially lower than the cost of aquiring OS licenses from Microsoft.
Cost of ownership is another issue, as Linux isn't as remotely upgradable yet as it needs to be for these uses, but that innovation is coming for both Microsoft and Linux in time.
Over all, looking at just the embedded device market, Linux presents a credible threat to Microsoft, sapping the growth rate needed out of this marketplace that would have gone to the Windows hegemony as Microsoft tried to leverage it's existing innovation.
Looking at the server market, it is more bleak for Microsoft. In short, Linux wins the applicability (due to customization capabilities - want a fast database server? Build the OS to specifically run the database). Linux loses the ecosystem argument for now, but ecosystems are far less important the more you move away from mass production markets; this one is shifting towards Linux rapidly. Linux wins the cost of aquisition aspect hands down and cost of ownership is being proven to be the Achille's heel of Microsoft.
If Linux saps away a significant portion of Microsoft's growth, what impact does that have on the company? Microsoft, even in this down economy has a P/E ratio of 51. This means that a tremendous expectation of earnings growth is built into the company's stock price and if that growth doesn't materialize, the stock is in jeopardy. Microsoft stock is priced in the market expecting a 30% growth rate in earnings. If their market growth is capped by competition, they will need to cut costs and raise prices in their existing markets to keep the stock price up, which will exacerbate the situation. In short, the embedded systems market and the server market represent two rocks, and their shareholder expectations are the proverbial hard place.
So what if the stock drops? Microsoft has underpaid it's employees by as much as 30% compared to market wages, compensating them with stock options. Lose the option value and the operating expense for the company goes up 17%, further depressing earnings, or they lose employees. The dastardly side of losing employees is what IBM learned - when a company is in trouble, the highly valued employees (ie. the ones that can get other employement quickly), scatter first, leaving the undesirables behind to screw things up.
Additionally, losing the stock value takes Microsoft's credit card away. Microsoft has, to a large extent, built it's new businesses through acquisition of other technology companies (webTV, Foxpro, Great Plains, etc.) and the ability to swallow new technologies on credit (stock given away in exchange for future accretive earnings) goes away, leaving them with the challenge of paying cash, which is abhored by Wall Street for a variety of reasons (screws with earnings, risk no longer tracks reward during the acquisition process, etc.)
So, where does this leave us? The PC market is a graveyard, software for consumers is relegated to games and utilities and the whole IT industry is in a slump. Microsoft is a big, fat juicy target for a lot of IT directors looking to cut costs and, as Geoffrey Moore pointed out, markets shift when the early majority customer base makes their move. Amazon is clearly not an early adopter or an innovator (in the sense of the Moore term). Where there is one early majority customer, there are typically many many others at work.
Want to beat Microsoft? Give up on the wasted energy behind creating a better desktop; there is no growth in that market to do any real damage to Microsoft. Instead, build a better Xbox with Linux, build a better PDA, build a better server (oops, already there) and sap the growth from the company. The efforts in the Linux community to innovate is best exerted in the direction of markets to come, not markets that are.
Technology Marketing is what happens when people turn their hard work over to people paid to manipulate others.
I did the math once now it time that GWB does the math, or can he?
Only 'flamers' flame!
To quote the linked article:
The online retailer spent $54 million on technology and content expenses in its third quarter, ended Sept. 30, compared with $71 million in the year-ago quarter, according to a filing with the Securities and Exchange Commission. The reduction was primarily because of Amazon's "migration to a Linux-based technology platform that utilizes a less-costly technology infrastructure, as well as general price reductions for data and telecommunication services due to market overcapacity," according to the statement. (Emphasis mine.)
So a lot of it was due to the move to Linux, but a significant portion of it was also due to their pipe(s) becoming much cheaper.
I don't think it's news to anyone that Linux can be used as a replacement for commercial Unix. It would be news if this article was about Amazon replacing Windows desktops with Linux (which it's not).
"But there are hidden costs to Linux, Microsoft argues. "I think a lot of customers are lured by the apparent low price of Linux," said Doug Miller, director of competitive strategy for Microsoft's Windows division. "They don't have a real issue with Linux, but it ends up costing them in the long run." Oh yes: 1. Linux as a core OS is EXTREMELY buggy and thus will require an expensive, disruptive OS upgrade when the next version comes around...or...not... 2. Linux doesn't adhere to open standards and thus in the long term its propietary standards require expensive custom integration products and a higher cost of systems integration...uhhmmmm.... 3. With Linux you have to disruptively upgrade to the next version when it comes out since the previous version will be rendered useless by the office software that runs on top of it....hmmm...nope... 4. With Linux, your systems will gradually become useless since linux system upgrades will demand hardware upgrades with each release, especially since the existing software is so buggy and the newer, fixed versions are only available for the new version. So your IT hardware budget increases...well... 5. With Linux, each OS upgrade the speed decreases or stays the same while the size bloats beyond recognition and useless features are lumped in that decrease stability and you have no choice but to include them....I thought I read that somewhere... Honestly, as a point of argument, can someone offer *SOME* rational devil's argument for this FUD comment?
Hey, I'm just your average shit and piss factory.
True people rent cars instead of buying them for several reasons. Usually because then they have a new car every 2 years and don't have to deal with fixing it when it starts to break down.
There is a problem with comparing cars to operating systems though. The bottom line is that not only is Linux free, but so are all of the upgrades, and you can upgrade to a brand new version over the net. Instead of having to go out and buy a new version and re-install their machine.
Using Windows XP is sort of like renting a car. Sure you get a new one every 1-2 years. But you have to go into the dealership to replace it and you are constantly paying. Linux would be a car that you don't pay for, it fixes itself and makes itself new everytime a new model is available.
Unfortunately, I agree that there are too many sheep out there though, and MS probably won't get tubed over this.
Disclamer - Opinion of Person
windoz 98 as a SERVER , please !, serving up what ?
I believe that Empty Sands was answering the point made by the AC. The AC stated that Linux was a poor desktop because it was "server capable".
This means that the system can run publicly accessable services such as http and ftp and a whole host of others. These services must be administered properly in order to remain secure and that is simply a headache not needed for a desktop system.
Empty Sands then replyed by pointing out that these services CAN be run under windows 9x/NT/2k desktops as well (I myself run an Apache-Tomcat server under Win98 in order to test my java servlets localy before uploading to the main test system, but my PC is firewalled off with zone alarm as an extra precaution), and that running these services, (although inefficent as far as high load would be concerned) is possible, and opens up the OS to the same security problems.
At least, that's how I read it.
"I'll take the red pill. No! Blue! AAAaaaahhhhhhhhh"
- Monty Python meets the Matrix