Market research firm Embedded Market Forecasters released a report today with data that indicates that embedded development projects using Microsoft’s “Windows Embedded” operating system platforms (including Windows CE .NET and Windows XP Embedded) is 43% faster and costs 68% less on average than projects using Embedded Linux. Further background, an executive summary, and a link for free downloading of the full report are available in this WindowsForDevices web site.
Report: Embedding Windows Costs Less than Embedding Linux
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
In unrelated news, it turns out that Embeded Market Forecasters is a wholy oned subsidiery of Microsoft. This, ofcourse, in no way affects their credibility.
research paid for by microsoft…. ’nuff said, really.
no comparison to other embedded os’s (QNX, etc.), of course, because microsoft is only worried about looking better than linux. increasingly so as they lost the Munich deal (along with several asian countries, it seems). granted those aren’t embedded deals, but they need to diseminate their microsoft-is-better-than-linux propaganda wherever they can…
because embedded Windows locks you up in your BMW extra-free all day long — That is *real* value Linux simply can’t beat — matter of fact…
I’m sure it runs find if you don’t mind your airplane having a 75% chance of not crashing and 50% chance of it landing on a runway. Oh, and you wouldn’t want to network it iether as that would inevitably be exorbatant with Windows as well as it would have vertually no security.
68% cheaper is only a dubious take on development costs. They coveniently left royalties out of the equation.
Yeah, but did they compare MS and Linux against NetBSD?
So Embedded Linux projects cost more and take longer. What kind of projects are they? Are we talking about a Linux encrypted cell fax vs a WinCE-GameBoy? Blanket statements that don’t address basic questions serve only to trash the credibility of whoever is making them (whoever they might be… why would they refuse to identified?)
I don’t doubt that Embedded Linux is complicated and doesn’t have as comprehensive a toolset available as Windows – regular desktop Linux doesn’t either. OTOH, it doesn’t crash or hand out your data to any teenager who wants it, and it’s at least possible to debug and fix OS problems.
Man I am really tired of companies like MS paying/bribing for “reports” just to put themselves in a positive light. Especially when we all know the real truth of the matter. We all know the real toll of the MS tax and it’s long term burden. Just like the reports that said that Linux has a higher TCO, ( which MS paid for ) I would take this report with a grain of salt. More marketing FUD is what this amounts to in the end.
For embedded devices, development costs are marginal. What counts is the manufacturing costs: this is where Windows CE licenses end up costing a lot more than Linux licenses. So, in other words, you pay a little more up front, but you end up making a lot more money. Why do you think the seven biggest Japanese CE companies teamed up to form CELF?
In other words, the study may or may not be accurate (it does say it was paid for by MicroSoft, so it might have been shelved if it hadn’t supported their view), but this is irrelevant, as development costs are often negligible as opposed to manufacturing (and thus, licensing) costs when it comes to “embedded” OSes in Consumer Electronics products.
The report was funded, in part, by Microsoft. “Microsoft paid me for my time to develop the framework and to analyze the data from a very reputable, very large, third party,” Krasner told WindowsForDevices.com editor Rick Lehrbaum.
That’s sorta an important point that should be in the article summary, dontcha think?
The reality is that very few companies are shipping millions of units. Most people are shipping 10-100 units. When you are in low volumes, development time and costs end up being far greater then runtime costs. And in high volume cases, runtime costs get very, very cheap. Often times there are one-time-buyouts for big projects. Cost of ownership issues are not as simple as people make them out to be.
Most people are shipping 10-100 units.
Well, it depends what kind of embedded devices we’re talking about: if it’s a new digital camera, GPS device and the like, you may be looking at hundreds of thousands of units (if not more). But it’s true that few companies can afford this – however they represent a very large portion of the market.
So, if this study is indeed accurate (as it’s been said, it’s being financed by MS), then we’ll end up with a paradoxical situation, where Windows CE is more advantageous for small companies who produce limited units, and Linux is more advantageous for large companies who produce very large amounts of units…not exactly the situation Microsoft wants, I imagine! (Since they’d make most of their money from the licensing fees on each individual units sold…)
I remember all that bullshit with microsoft sponsored “researches” like .net “pet store” vs. Java,”lower” windows tco and so on – I know that FUD strategy,but there are the facts: most enterprise apps are still Java,and adoption of .net is slow and below all expectations,Munchen chose SuSE,Telstra is switching to Linux – ah,now embedded windows!
… isn’t there some kind of law or legal precident that makes this kind of thing illegal? I mean, they were blasted for creating a fake “grass roots movement” supporting them, but in the end, the result was what? Do consumers trust them less? Do consumers take what they say with a grain of salt? I don’t really think there’s any balancing force going on. Exposing them seems to be only to the benefit of people like us OS geeks while having no lasting effect on their future PR.
From most of the situation I’ve seen that’s the case. It’s usually cheaper for a small company to go with Microsoft products, since they have lots of off the shelf solutions and are fairly easy to figure out how to use. Unices and the wannabes, in general make more sense for large companies who can afford to spend the time tailoring the system to their needs.
It’s a comparison between a tailored suit and one off the rack.
Personally I’d be more inclined to use a purpose built OS to either one of those, ex. PalmOS.
It might be 100% legal, but legal != ethical.
So you’re linking to a report that says embedded windows is cheaper than embedded linux, and this report is on windowsfordevices.com
Wow.
In related news, over on bmw.com, they say their cars are the ultimate driving machines. But yet, on pontiac.com, they’re building driving excitement.
well, Sharp said that the total cost of using Linux on the Zaurus was much higher then if they had used WinCE. I really think it needs to get into millions of units before the trade-off is compensated.
Sharp is not a good example because they developed a platform based on Linux akin to Windows CE. Perhaps you could compare the cost of developing the Zaurus platform to the cost that M$ spent to develop Win CE platform. Otherwise, you are comparing apples and oranges.
Isn’t that the guy who wrote in a presentation:
There are important post-September 11 issues that will drive the New Economy. These will effect the world’s economy  not just the embedded world
… shortly followed by ….
Post – September 11 driving forces that should be of interest and concern to embedded vendors:
* The likely demise of OPEC
* The fear of US reprisals
* RBOCs forced to open Broadband
http://216.239.51.104/search?q=cache:22EMiADRB5gJ:www.busandboard.c…
OK. Let’s see what else this guy has found.
Embedded developers were asked to identify the biggest problems they confront in hardware-software integration. Fifteen suggested answers were presented (plus “other”) and the respondent was limited to selecting 4 responses. This provided a preferential ranking by order of importance. The following items were the most frequently selected problems.
Limited Visibility into the Complete System (65%)
Limited Ability to Trace (54%)
Limited Ability to Control Execution (42%)
Significant Intrusiveness Into the System (40%)
http://216.239.51.104/search?q=cache:m6TJXiPqwFAJ:www.embedded-fore…
So let me get this straight. This guy, who is the kind of guy who predicts the demise of OPEC and then attempts to make that on topic w.r.t embedded systems, has performed research that shows that 3 of the top 4 problems embedded developers face are essentially solved by having access to all the source code. He then manages, after getting paid by Microsoft, to conclude that actually Windows is better after all.
Here’s an idea. Why don’t the analysts give the engineers some credit, and let them get on with their job, using whatever tools they find best to work with? Radical I know.
Okay – so what is a fare comparison then? The thing is that they just simply worked with Qt and Lineo to do the embedded work. Just as HP works with various companies in tawian to do thier WinCE work. The reality is that it cost SHARP, and they are using an open platform that others are now using – just like people using WinCE. But it costs more to use that platform then WinCE.
Linux is an option, but it isn’t free ($$ wise).
XP embedded is better than Linex in this configuration: an embedded application with graphical frontend on 800M CPU, 256M fresh memory. Linux can fit comfortably in 256M, but what about X-window?
I don’t think this follows the Scientific Method or Logic. There are obviously other factors that contribute to the cost of Embedding. Also, there are other factors that are not included beyond cost: ROI return on investment, features, and upgradability (2.6???).
Also, 100 companies does not seem like a large enough group. We also don’t know how the companies were chosen. They could have selected 50 competent or at least well respected companies (which were using Windows) while the linux embedding companies were startups or known screw ups (Enron, pets.com, JeneaneGarafolo Inc.)
There was also no control group or other checks on variables. All 100 companies could have been given both Linux and Windows for the same tasks and see which project came out better. It could very well be Apples and Oranges here. I’m willing to concede the fact that Windows could be better on some type of projects since this is all this test could have used. So, I’m not convinced here by this report. While you should always use the right tool for the right job, this might be a bad setup. Remember the studies that say Second Hand Smoke kills? Well, there’s studies that dispute that. At least one of them is wrong and we can at least conclude there are not enough studies about this. (Second hand windows could get you audited….)
Then why isn’t anyone embedding it? I can count 30 devices that run on EL off of the top of my head, and *2* that use MS technology. Not to mention the consortium of MAJORS that created the Embedded Linux group last week. LOL
“Microsoft admits critical flaw in nearly all Windows software
TED BRIDIS, AP Technology Writer Wednesday, July 16, 2003
(07-16) 14:25 PDT WASHINGTON (AP) —
Microsoft Corp. acknowledged a critical vulnerability Wednesday in nearly all versions of its flagship Windows operating system software, the first such design flaw to affect its latest Windows Server 2003 software.”
http://www.sfgate.com/cgi-bin/article.cgi?file=/news/archive/2003/0…
Yup these guys are going to protect us ! I think not !
“The announcement came one day after the Department of Homeland Security announced that it awarded a five-year, $90-million contract for Microsoft to supply all its most important desktop and server software for about 140,000 computers inside the new federal agency.”
http://www.securityfocus.com/news/6397
http://www.windowsfordevices.com
Nuff’ Said.
On an unrelated note http://www.scofordevices.com has a full report that SCO Unix Cost less than any other OS on earth, and does not have IP ‘issues’.
I ran out of fingers to count all the ad hominems…
Those Linux fanboys out there who failed to actually criticise the methodology (as opposed to shooting the messenger) deserve a good spanking.
We don’t anything about the sample, about the scope of each project, about all the costs included, about licensing fees on the WinCE side, etc., etc., etc.
Therefore, one can’t form any conclusion as to whether the study is valid — although I do not doubt the sincerity of the author.
It’s curious to me that there was no mention of what I think is the number one Linux embedded success story — Tivo.
In any event, Linux embedded people should work on creating better and easier to use tools.
Since when was pre-paid (hence result pre-determined) ‘report’ a methodology?
“Since when was pre-paid (hence result pre-determined) ‘report’ a methodology?”
Perhaps you didn’t read what I wrote the first time around. I made one reference to ad hominems, and another to shooting the messenger.
To say that because Microsoft paid for a report means that it is meaningless is not a rational argument. It is argumentum ad hominem by definition.
At worst what Microsoft’s funding means is that people analysing the report should pay extra careful attention to the way in which the research was carried out. Simply receiving funding does not instantly mean the conclusion of the report is invalid – it is quite irrational to argue otherwise.
What needs to be recognised is that Microsoft is in business to make money, and when they believe that they have an advantage over the competition, they will want to tell the world about it. Paying someone to do a report on the issue is not an unusual practice.
All the ad hominem spinners also seem happy to criticise the professional integrity of anyone who takes it upon themself to carry out such research. If one does a Google on the man who authored the report, I would imagine that his credentials are far above those of the majority of respondants to this article, and particularly those who dish out ad hominem attacks against his report.
Criticism of the report is fine, but criticising it just because Microsoft funded it is irrational. Try http://www.dictionary.com for “ad hominem” for a definition.
Man, I’m so tired about ABMers always slandering Microsoft for no good reason…
“What needs to be recognised is that Microsoft is in business to make money, and when they believe that they have an advantage over the competition, they will want to tell the world about it. Paying someone to do a report on the issue is not an unusual practice. ”
It’s the how. not the what that gets people’s goat. Saying that “one’s in business to make money” and leaving it at that is like saying that the mob is in business to make money and not looking at the rest of the picture.
“At worst what Microsoft’s funding means is that people analysing the report should pay extra careful attention to the way in which the research was carried out. Simply receiving funding does not instantly mean the conclusion of the report is invalid – it is quite irrational to argue otherwise. ”
And likewise it’s irresponsible to not look at the track records of these “reports”. If Microsoft wants people to trust their reports they are going to have to earn it. Not give it away like some cheap carnival prize.
“And likewise it’s irresponsible to not look at the track records of these “reports””
This is another ad hominem.
Which is more rational:
a) Consider a report based on what is actually says, rather than who funded it (keeping in mind that a fairly well known analyst is putting his credibility on the line by openly acknowledging that Microsoft paid for the report)
b) Write the report off without consideration just because of who funded it (and your lack of trust in that source).
As for me, I’ll pick a).
As for history, what ever happened to the one from … mindcraft was it? … where NT was shown to out perform Linux (this was back in 99 I think, so it’s an out of date finding now). The Linux camp kicked and screamed, so Mindcraft did an open benchmark and the results came back showing virtually the same conclusions as the original research. The subsequent silence from the Linux camp following was quite deafening, if I remember correctly (maybe someone can show how the Open benchmark was still biased, I may have missed it).
Regardless, history is no reason to write off a report. It should certainly make you suspicious, but rational criticisms can only be based on addressing the methods used to perform the research.
Close your eyes and dream for a moment: (okay, better open them again to read)
Let’s just imagine a scenario where Microsoft did pay for a report, but despite Microsoft’s funding, the research was not at all biased, and everything it said was accurate. This is just a hypothetical situation remember. According to what some of you are saying, that report should be thrown in the trash (inspite of it’s accuracy) just because Microsoft funded it.
This isn’t rational. The report should be based on the merits of the report, and the merits of the report alone. If the report is full of flaws, then you have every right to jump on the bandwagon and accuse Microsoft of propaganda, but a bunch of you are putting the X Window System before the boot loader.
My study shows that the cost of developing embedded application under embedded Linux is 78.5% faster and 83% cheaper than under any Windows product
So I welcome the comment/bashing from the community on the above phrase
“‘And likewise it’s irresponsible to not look at the track records of these “reports’
This is another ad hominem.”
Uh I know you like to hear yourself say “ad hominem”, but it’s not. If I had said that that’s the only basis to accepting, or rejecting the report then you would have had your “ad hominem”
Big hints David:”Likewise” and “Reputation”.
Even if everything the article said was true, consider where it came from. I don’t go to Linux.org to find out how cost effective Windows is. The article MIGHT be full of truth, but a M$ paid article from a pro-M$ site is not going to carry a lot of water.
M$ already knows they lost. Whether to Linux or a BSD, it’s happened. They can stretch it out all they want, won’t change facts. Using money to create their own ‘fanboys’ won’t sway the crowd. I don’t understand how anyone can support a criminal organisation (M$) and still sleep at night.
Concerning the MindCraft fiasco…
http://www.opensource.org/halloween/halloween6.php
Dave, your precious M$ destroyed this company’s credibility. And you quote it like it means something. Without even understanding what happened.
“Regardless, history is no reason to write off a report. It should certainly make you suspicious, but rational criticisms can only be based on addressing the methods used to perform the research.” — Agreed, but Dave, lets try arguing on facts from now on.
wow…any proof, or you just barking your bullshit..
“Oh, and you wouldn’t want to network it iether as that would inevitably be exorbatant with Windows as well as it would have vertually no security.”
That’s not really a flaw in Windows – it just shows an incompetent admin. An machine is only as secure as the services running on it. It’s not that difficult to disable all the services that aren’t required, and replace insecure ones with secure equivalents.
Isn’t making a fortune while other people die of hunger a crime in the first place? Or am i attacking the fundaments of the American society with this?
This is not only a MS crime but of our whole society, everbody is affraid they get less and let a child die because of there greed. I am guilty of that too!
As for MS, they deserve to get more competition, this is better for the users, developers and the oem guys.
This is where there crime is begining, to sort off, forces
oem companies to ship their Windows and then give them a better price. I can not (in general) buy a Laptop without Windows on it, even if i do not want it. This is because MS
forces them and that is maybe a crime..
Microsoft is becoming more and more like SUN Microsystems, and SUN Microsystems is focusing more and more on their products.
Microsofts enemy isn’t Linux but themselves. Instead of bad mouthing Linux, how about a technique called “positive promotion”. Instead of sledging the competition why no simply promote the things that are great about their products.
Or would that just be too logical?
All WinCE Users purchased a Licence with their product. This means a cost surplus. On the long run, a Linux strategy could mean lower prices which means more devices sold. The vendor generates no profit from the WinCE-license, but from the quantity of sold devices…
After what i heard Embedded Linux actually isn’t so good.
(correct me if you KNOW that i’m wrong)
Not comparing to Windows but compared to OS’es that are made for embedded devices like QNX, vxWorks, NetBSD and TRON
Theese(with exception of NetBSD) are systems designed to work with embedded devices from the beginning.
All this probably counts to Windows OS’s to.
And that report is probalby another of thoose paid-by-ms “objective” reports…
XP embedded is better than Linex in this configuration: an embedded application with graphical frontend on 800M CPU, 256M fresh memory. Linux can fit comfortably in 256M, but what about X-window?
It’s _very_ easy to embed Linux on limited space. We run our product on a 64MB flash where Linux (system + xfree86 + Mesa) takes up half of the space.
IMHO, for embedded systems, having open source OS is almost crytical. When you sell the box to a customer, and then critical bug in OS is discovered, and your OS vendor promises to fix it in next release in about 6 month, because you too small customer – you are in trouble. If the OS is open source, you can at least try to fix it yourself. That’s why I always prefer open source embedded OSs – Linux, ThreadX, BSD – over any closed source.
“M$ already knows they lost. ”
Yeah, right. Wishful thinking. It’s amazing how some people completely lost their touch with reality.
Which is more rational:
a) Consider a report that claims that tobacco smoke is good for you, written by a group of respected scientists, funded by the tabacco industry
b) Write the report off without consideration just because of the conflict of interest between the report and who wrote it.
As for David, he smokes 25 packs of cigarettes a day, feeling completely justified! 😉
As for the rest of us, we understand the terms “conflict of interest” (look it up in http://www.dictionary.com) and “means, motive and opportunity” (categories used to establish guilt). You can spin all you want, but it is undeniable that both terms fit this scenario!
great..my aunt has wince(inpain) on her new bmw 74o.. the car refuses to function properly at times.. for no apparent reason. Why didnt the author throw in the comparison to qnx and tron?? (Clue: he didnt want to compare to any more “Real” os)
>Yeah, right. Wishful thinking. It’s amazing how some people
>completely lost their touch with reality.
Yeah! How are we going to get you back in to the reality?
I would take the red pil if i where you..
So, when MS has 95% of OS market and 40 Billion in cash, who’s ignoring the facts, you or me ? Get real. Comments like “M$ already knows they lost”, ignoring the “M$”, “Windoze” etc. are at least childish.
I would take the purple pill if I were you.
Finally M$ is tapping some humor outta windoze box’s. This is superbe jiddish humor. My goodness rotfl. And jiddish humor is great.
This also leads to a 43% greater chance that M$ will blatently steal your IP and undercut you by producing it in a third world country. There is also a 68%+ chance that the Court of Appeals for the Federal Circuit will do nothing about it. Woohoo.
Yes, it is:
http://usvms.gpo.gov/ms-final2.html
AZ
Look, I hate Microsoft, but it’s like in a fun way. It’s like how you boo Rowdy Roddy Piper and cheer for Hulk Hogan. To say it’s a crime to make money while people starve is INSANE!!! Didn’t Bill gates donate like $100 Million for AIDS research? With your logic I guess we should all starve together. How about this, it’s a crime to protest eating animals while people starve. Off topic.
Point is, Bill Gates is evil, but not as evil as Saddam Hussein. Bill Gates doesn’t use nerve gas and torture to keep in power. Yes, Gates is mean, but he’s done a pittance compared to enron, worldcom, and JeaneaneGarafolo Inc. Nobody died because people are running Windows. I think. The world would be a better place if more people used Linux, and that’s my goal, but stop blaming Bill Gates for Wolrd Poverty. He’s a jerk. He basically pioneered Napster for his software (well sort of). We boo him, but he’s not the Anti-Christ. Maybe the Anti-Moses, but not the Anti-Christ.
If we want Linux to become “successful” or whatever our Goal is for linux, IT’S UP TO US. We can’t complain. We dont’ get Affirmative Action for Linux. We don’t get force other people to use Linux. We cannot complain about a Vast Right Beam Conspiracy stopping Windows. The White man is not trying keep the LInux Brothers down.
Now if you want a really interesting Idea, I say we teach monks to code and have them do the development, like how they used to copy books in the Dark Ages….
From the bottom of the article page:
“Dr. Jerry Krasner, an authority on the embedded marketplace, was paid by Microsoft to develop a cost estimation framework for evaluating embedded development platforms.”
It’s _very_ easy to embed Linux on limited space. We run our product on a 64MB flash where Linux (system + xfree86 + Mesa) takes up half of the space.
Yeah, you da man! Big deal, I say. You can fit Windows XP Embedded in half than that
IMHO, for embedded systems, having open source OS is almost crytical
Windows CE comes with sources. Enough customers asked for them (for the reasons you outlined) to make Microsoft give up, and give away the sources for free
David, “sponsored” research means it’s not trustworthy, it may be an interesting exercise to check the methodology, but the conclusion should be thrown in the bin for obvious reasons.
If you don’t believe this, I suggest you pick up smoking. It’s good for you (really : http://193.78.190.200/wsj/wsjcom__does_passive_smoke_kill.html ), and it’ll cut a few years of your life, so we don’t have to hear this again.
snowdog, your little reference at opensource.org failed to mention the subsequent open benchmark which Mindcraft performed, and which I refered to previously.
As I said, the Linux crowd kicked and screamed, so Mindcraft performed an open benchmark, revalidating their study, and the Linux crowd now seems to have conveniently forgotten, if your little “historical” record which skips details is anything to go by. Why did your little article ignore the subsequent open benchmark? Why were Linux supporters not keen to stand up and prove their beloved OS better than NT?
As for those smoking reports, those were famously wrong, but what does that have to do with Microsoft? If a black man (or green, or any other colour) commits a crime do you assume that all black men are criminals? Maybe you do, but I don’t.
It is irrational to assume that Microsoft funded reports are absolutely 100% certain to be biased and not worth reading because the smoking industry funded reports which were inaccurate.
Whether you like it or not, these reports *may* present perfectly valid research, and it is irrational to argue otherwise.