Slashdot Mirror
Open Source & Embedded
Embedded Geek writes "Jack Ganssle has posted a column at embedded.com pondering whether Red Hat and other open source companies serving the embedded community are due a shakeout similar to the dot com collapse. He cites Red Hat's March cuts in their embedded division and their losses of $80M to $140M a year. He admits, though, that because the embedded market is smaller and many companies are privately held it is difficult to get a pulse on what's going on behind closed doors.
"
I thought RedHat made like $7.50 last year, or something! Now it's losing 150 million?! Quite a turnaround. :(
Right now, it's more like "Which company isn't having financial issues."
This stuff is getting rather old. Everyone is having a tough time being profitable and I don't believe "opensource" is the reason why RedHat is struggling. You name the company, products and services that were selling well in 97-00 are no longer doing that well. EMC and Sun which used to be very profitable at what they did are having to switch gears. Cisco is having it's share of problems. Software or hardware everyone is struggling and Opensource companies are still trying to figure out how to make a profit on a service model.
Folks this isn't news and really doesn't need to be argued over but Opensource isn't the next "Dot com bubble" that is going to burst. Opensource, Redhat, whatever... everything is having trouble right now. Let's not make this out to be more than it really is.
Open Source companies aren't really built on the hype and hyperbole of the DotComs. Having a substantial product allows for more longevity (sp?) than a web based retailer.
Think nothing is impossible? Try slamming a revolving door.
It is easier to offload failing operations in a market upturn than a downturn. At leats then somebody might pay you something for your company.
The sad reality is that - as the economy picks up speed - the number of failures/fire sales will also accelerate, at least in the short term.
I'm not saying RH is going to go - I doubt it - but a lot of "big players" could end up being sold for very little.
I agree that people will try alternatives. But to be realistic, what alternatives are there really?
How many open source project does it exists thats really finished? It like one long beta period, and instead of getting things finished people implements new features instead.
Sad, but true.
I don't think many Open Source projects will ever be considered "finished," because that is exactly the point - the source is open - thus allowing for additions/changes, and it will never be "finished".
And anyways, just what exactly does "finished" mean when talking about OSes? From the way I see it, a complete operating system would be a very difficult thing to "finish," especially an open sourced one where many, many people around the globe contribute.
One problem with embedded Linux solutions is that developers are compelled to release their modifications as per the GPL. Whether or not this is a good thing (and there are some very good reasons for both GPL-style and closed-source software, beyond the scope of this comment), some companies just don't enjoy the thought of having to release the source code for their embedded products.
Embedded BSD is the solution to that problem, and in fact various forms of embedded BSD have been around for a long time and are going strong. Perhaps if Linux changed its license from GPL to LGPL it might help in this situation, to gain more acceptance from the business community.
Cedric Balthazar Rotherwood
Sun Certified Programmer for the Java Platform +
System Admin. for Solaris
Everyone having problems now? No, microsoft is making profit, oracle also makes profit.
Microsoft and Oracle are everybody? What he means is the economy in general, as well as most of the tech industry is experiencing a down turn. Providing exceptions doesn't change the stats quo that is happening at the moment
Think nothing is impossible? Try slamming a revolving door.
As far as I know, Cygnus was profitable, at least in the Golden Years just before Open Source more or less hit the mainstream press. There is an interview at developerWorks in which Cygnus Solutions cofounder Michael Tiemann claims that they were profitable until the venture capitalists came, since investors "give you money so you can actually accelerate the rate of spending versus revenue".
It's probably not too easy to turn back the clock for Red Hat, as there is increasing competition on the GCC customization market, where companies pay immense sums for adaption of GCC to certain microprocessor platforms and support for that GCC derivate (at least they paid these sums when Cygnus didn't have much competition!).
It like one long beta period, and instead of getting things finished people implements new features instead.
Sounds like a certain operating system I read something about.
I've had enough abrasive sigs. Kittens are cute and fuzzy.
Red Hat's March cuts in their embedded division and their losses of $80M to $140M a year.
Do I need to explain the concept of "Goodwill" yet again? Why do people keep this shit up? Are they holding short positions in RHAT?
Goodwill is the amount of money a company pays for another company, that is above the value of the tangible assets of the aquired company. This may include brand names, patents, and other intangibles. If a company buys a company that turns out to be overpriced later on, then the buying company will have lots and lots of goodwill that must eventually be charged against earnings.
Red Hat did not lose those large amounts of "real assets", rather, almost all of that "loss" was a write down of goodwill from previous aqusitions.
Look at it this way.
1) RHAT IPOs
2) RHAT stock becomes grossly overvalued
3) RHAT makes a secondary offering, cashing in on their grossly overvalued stock in a big way.
4) RHAT goes on a buying spree, spending their money that they got for free, buying companies like Cygnus.
5) Cygnus was also pretty overvalued, so a lot of goodwill ends up in the "assets" column of RHATs balance sheet.
6) RHAT has to write off chunks of this goodwill against earnings later on.
The key is that RHAT got this money for free. Had they not made the secondary offering, they wouldn't have had the money to make the aqusitions in the first place. Sure, that money ultimately came from idiot investors that paid $200 a share for RHAT, but it didn't come from any direct business or financing activities that had an opportunity cost for RHAT (such as debt financing).
So no, they aren't losing those staggering numbers each quarter, in fact they are breaking even for the last 3 quarters or so.
Goowill can be abused. "One-time-charges" can be abused (see Cisco writing off billions of real, tangible, inventory).... but in this case... there is no money lost, just worthless monopoly money that no longer exists, and hasn't existed for years, subtracted from a column on a balance sheet.
I've had enough abrasive sigs. Kittens are cute and fuzzy.
There is a saying the the software development industry and that is when you have done 90%, you have 90% to go. This means that when you have implemented 90% of the functionality you still have 90% of the job left to do. There is much truth in this.
This is the problem with almost all open source software, noone is interrested in really make a FINISHED product thats usable. It's time-consuming, boring and hard to finish a product and most people aren't very interrested in beeing bored on their free time.
Stuck in a small box
CPU crunches numbers:
Use Redhat or not?
I work for an (open source technologies) embedded development consultancy company. The only thing limiting our growth is finding competent personnel. Demand exceeds what we can deliver by a large factor. Linux has taken the embedded development world like a tornado grabbing a trailer. I see no bubble building up here.
Of course, other companies may have been less realistic in their growth perspectives and may have had more venture capital to burn.
In making a business plan, you will need to figure out what your differentiators and profit centers will be. It's obviously not going to be software sales - the folks who tried a royalty-based embedded system based on Linux went out of business faster, because they had free competition. In general it will be consulting services, and this is going to be a difficult business in a slow economy as businesses will try to do more with their own engineers. Businesses also have an incentive to use their own engineers for embedded work, as they don't want to be in the situation of losing the recipie for one of their own products. That can happen more easily when an outside vendor does the work.
I created the user-mode half of most Linux embedded systems - a program called busybox. It's everywhere. I used the GPL. Because of that decision, the person who put the most effort into maintaining that program, after me, is still working on it and is able to offer his consulting services on it. Had I not used the GPL, he would have had to give it up when his previous employer was one of those shaken out. I have a lot more sympathy for him than the employer. Also, had I used a license other than the GPL, the program would not have become an open standard for embedded - everyone would have been making their own proprietary additions rather than cooperating. And I didn't care that companies could not lock in a revenue capture on busybox - why should I?
I think Debian has the best "business plan" of any Linux distribution: don't even try to make money. The people who use Debian as a cost center (HP, for example) pay for its development, and they are very clear about what their profit centers are. This is why I think Debian will eventually end up on top.
Bruce
Bruce Perens.
Bruce
Bruce Perens.
Red Hat's embedded division wasn't selling Linux based solutions, it was selling the eCos product that came from its acquisition of Cygnus. A minimal eCos kernal can be about 4K of ROM and 1K of RAM. The kernel isn't a POSIX compliant API, (although they seem to have recently built POSIX/Linux compatibility layer on it called EL/IX which would put it in the same ballpark as QNX.) I agree with your main point though, many of the big system technologies (C++, Java, Linux) seem out of place for embedded system design.
Karma Whorin' Galor'n
NoMoreNicksLeft
"When will he stop?" Pon-
dered the troll
"Super-slash-poetry
doesn't amuse me; my
super-slash-trollery's
withered my soul!"
surrreee they dont.
so.. does your business have a yearly contract with MS? Sun? IBM?
right. thought so.
... hi bingo
I agree that open source is a shared cost centre. One of the "disruptive" things about it is that small companies can benefit to the same degree as large companies.
In effect, this lowers the barriers to entry. Could a small company fund the development of an operating system? No. But anyone can take a Linux distro...
So does Anonymous Coward have good karma?
However, just so people don't think it's just a Linux thing... Annasoft, a leading Microsoft embedded partner, just died.
Let's see...
You'd be surprised at how many devices have embedded firmware.
Lets see... microcontrollers used in LOTS of consumer devices (toys, microwaves, cd players, radios, etc.), adjunct devices (for example, the ethernet palm pilot cradle I worked on a while back), barcode scanners, all manner of microcontrollers/firmware on computer peripherals (PCI cards, printers, scanners, etc.), cordless phones (do you think you get the caller ID message without firmware?) etc. etc.
WindRiver is the biggest embedded systems OS provider, selling VxWorks.
You pay a one time fee for Tornado seats. What does it buy you? Some packaged up free software compilers (gcc!) and limited tech support. Same as Red Hat.
Outside of the support contract, you can get someone on-site from WindRiver for $2500 a day. And believe me, at some point, you will pay it because you've got to get your product shipping... sometimes you're willing to pay almost anything.
What makes them different? Per-copy license fees for VxWorks when you actually ship your product. And they ain't cheap. So guess what? There's a definite advantage to Open Source. No license fees! And the code works just as good as the closed source stuff, AND you don't have to beg WindRiver support for source code when things don't work right (AND this DOES happen). That's the customer perspective.
From Open Source perspective, one can make plenty of money just on support, just like WindRiver does. Adopt some open embedded code, learn it inside and out and start charging $100 an hour to support it. You can't help but be profitable as long as there are enough contracts around.
Linux has taken a huge percentage of the "do it yourself" market which was actually about 60% of embedded. Of commercial embedded development, WindRiver used to own about 60%, but that number is going down fast. Linux is expected to win more new design wins in the next year than WindRiver, making it the #1 choice for new designs. An important thing to remember in embedded is that a large percentage of the market has legacy deployments in the market already and those will continue to be VxWorks (or pSOS) for the forceable future. But Linux is here to stay...
Of course, I do have my doubts that a company like Red Hat makes sense. Open source is best for specialized, highly-skilled consultants and professionals. It also makes sense for a few hardware vendors and large, established consulting shops like IBM. But companies like Red Hat and VA Linux aren't in that league.
I haven't got any spam trying to sell me embedded engineering manpower yet, like I got several for web design, Virusal Basic and other markets that have crashed recently.
I wonder, though, why companies aren't giving more money to the people whose software they use. Has Bruce Perens ever received a single penny for busybox? From any of the companies that use it? I don't know, but it certainly hasn't happened for me. It's sad that companies using open source software usually only spend money on training people who never did any embedded stuff rather than on the people whose ready-made software they are going to rip.
eCOS, BSD, and RT Linux have established a clear and viable alternative to proprietary solutions, and they are expanding the extent to which embedded systems are becoming source-code compatible with one another.
For medium-to-large systems that could actually take advantage of 32-bit open source solutions, suddenly QNX, pSOS, and the others are talking price like never before. I'm sorry if Red Hat is losing money, but at least they've established a market that stops the established vendors from locking you in and gouging you.
Open source solutions have created an environment where the choice is no longer which vendor to be locked-in to because now there is the none-of-the-above choice. So now if you go with QNX because it is indeed the best option, they still have to be nice to you because you can walk. Everybody wins.
So the point is that the price of all embedded kernel source has equalized, proprietary or otherwise, the vendors now all compete on service, turn-key solutions, and technical merit.
You'd be surprised at how many devices have embedded firmware. ... toys, microwaves, cd players, radios, etc., ... ethernet palm pilot cradle I worked on a while back ... barcode scanners ... all manner of microcontrollers/firmware on computer peripherals (PCI cards, printers, scanners, etc.), cordless phones (do you think you get the caller ID message without firmware?) etc. etc.
And don't think for a minute that a single one of those devices runs anything that even remotely resembles a general-purpose operating system.
Companies need long-term planning for their investments. Personally, I'd rather pay $1000 for software knowing that it goes to a company that is most likely going to be supporting the software for the next 2 years, than $500 for a package where this is not likely at all. Unfortunately, only very few O.S. projects come with some sort of established development process that guaratees that they're going to be "supported" (bug fixes...) in the next 2 years.
Also, the best Open Source programs are so popular and widespread that you get your support for them elsewhere (contracting...), so money spent by companies "on" these products goes to those other people. Finally, don't expect a company to donate money if it doesn't have some visible positive effect on the company itself ...
IMHO, Open Source developers need to find business models that work well together with their products, e.g. web server software developers re-selling SSL certificates (I don't care where I get mine, and I don't mind paying slightly more if the money goes to the right people), compiler tools programmers offering affiliate links to recommended books about programming languages etc. ... It's much easier to say "thank you" by chosing you as the reseller of something (not merchandise, unless it's something really useful!) you're going to buy anyway at more or less the same price, than by donating money with nothing in return.
"I love my job, but I hate talking to people like you" (Freddie Mercury)
Re-reading your comment, I noticed that you actually meant companies who included Open Source software in their commercial software packages... My rant wasn't specific to that situation, but to the use (not commercial re-distribution) of Open Source software by companies in general. So, it's just a rant ;-)
"I love my job, but I hate talking to people like you" (Freddie Mercury)
Why can't they stay in their shitty third world counties ...
Which county are you from? Hazzard County??
cpeterso
Credit where credit is due.
Erik Anderson maintains the busybox todoy.
The employer you have spoken of is/was Lineo.
--
Karma 50, and all I got was this lousy T-Shirt.
Wasabi Systems, an embedded NetBSD company, is doing really well, last time I heard.
The eCos kernel, the kernel that RedHat acquired when it bought Cygnus, does not have a POSIX compatible, or even unix like, kernel interface. It has a kernel interface, it just looks very different than Unix. They weren't creating an incompatible kernel just to be different, they were doing it to create a kernel with features ideally suited for the development of embedded devices.
As an example, a separate thread of execution is not started with fork() or pthread_create(), but rather a thread is created with cyg_thread_create() or sta_tsk() if the ITRON API is built into the kernel. If you look at the eCos API and something with a Unix or POSIX API you will notice that few if any of the same calls exist on the two platforms. Just to reiterate, so I'm not confusing you again. RedHat's eCos is a kernel designed for embedded systems and is not at all similar to the POSIX standard.
Now QNX as a developer of of software for the embedded market took an interesting strategy. Since there is so much software written for a POSIX style kernel, and it is an API that many people are familiar with, they created a kernel small enough for embedded devices, but gave it a POSIX API. This may make some tradeoffs in many directions, since what the program needs from the kernel and what the kernel needs to inform the code is very different for embedded devices than for a desktop or server OS.
Now to my comparison between eCos and QNX. When eCos is built with its optional EL/IX component, it starts to implement a subset of the POSIX API. So calls like read(), write(), and fcntl() are available. If you flip on the EL/IX switch, it starts getting a lot more familiar to Unix developers, which getting the product a little towards what QNX's product line does. Or at least having the same advantages.
Unfortunately, I can't quite understand how you read what I wrote originally and picked out the two sentence fragments you quoted as being a summary. If you could point out where my original post was unclear, I can try to avoid making the same mistake in the future.
Finally, to directly answer your points.