Article: The Business of Embedded Linux
May 24, 2002 — by LinuxDevices Staff — from the LinuxDevices Archive — 2 viewsBigger pie, too many diners?
If Inder Singh, chairman of the Embedded Linux Consortium (ELC) and CEO and Chairman of LynuxWorks, is to be believed, Linux is on the verge of becoming the real… world standard for most embedded devices. But, technology alone can't win you commercial success. No, not even in the embedded space. For, as Dan Kusnetzky, IDC's vice president of system software research, likes to say, you're not just selling to Dilbert, you're also selling to Dilbert's boss.
So the marketing, management, and sales side — the business side — of embedded Linux companies plays as large a role as the technology side in embedded Linux success. Many people don't want to hear this, but no less a figure than Bob Young, one of Red Hat's founders and currently a director, has said that “Good management with crappy products will beat crappy management with good products every time.”
Best of all, of course, are good companies with good products. And the future for such companies looks bright. Stephen Balacco, Venture Development Corp. (VDC) Senior Analyst, concluded from VDC's recently completed research study, Linux's Future in the Embedded Systems Market, that “shipments will reach $306.6 million” by 2005.
Unfortunately, while that's a big pie, embedded Linux companies have a major business problem: it's hard to tell them apart. Tom Williams, Editor in Chief of RTC Magazine, says: “Everyone is trying to do the same business model, and the jury is still out on how well they're going work. All of these are small companies; they're basically trying to sell the same thing, and they can only compete on margins with big players like IBM.” Specifically, “They're all selling the same thing, and they can only differentiate between each other by support, pricing, and proprietary tools.”
Another problem is that some observers simply don't think that services and support alone is enough to keep an embedded company alive. As Victor Yodaiken, CEO of FSMLabs said in an interview with EETimes, “Free software has to be packaged with a product that you can sell, it's very difficult to do a pure, free-software model in the embedded space.”
Williams continues, “I think that companies that incorporate embedded Linux in their products will do well, but I don't see a big future in selling embedded Linux. If a company grows to a certain size, they can bring in their own embedded Linux team, so the customers for these embedded Linux companies will be small.”
In other words, the embedded Linux companies face the eternal problem of whether a company will outsource their embedded needs to them, or take them in-house with their own development staff. Sound unlikely? Think again. In Chris Lanfear and Balacco's VDC whitepaper entitled Linux's future in the embedded systems market, the second most popular choice for customers embedded Linux needs was, “roll-your-own.” So it is that Williams and many other observers see consolidation ahead for the embedded Linux companies.
It's hard to argue with that vision of the future. With no fewer than seven significant players — Lineo, LynuxWorks, MontaVista, Red Hat, REDSonic, RidgeRun, and TimeSys — the embedded Linux players outnumber the major business Linux distributors: Caldera, Red Hat, SuSE, and Turbolinux. And, of course, there are smaller embedded Linux companies as well. Even with a $306 million pie, it's hard to see all of them getting enough to survive.
Drilling deeper
At first, Lineo was the poster child for embedded Linux. Now, financially speaking, they're a red-haired stepchild. Lineo dreamed big — CEO Matt Harris admits saying that they had expected the embedded Linux market to be double its current size. Their dream has turned into something of a nightmare, with massive layoffs, bounced payroll checks, and an almost complete turnover of the executive suite since November 2001. Still, their products are well received, they've continued to make big deals, and The Canopy Group is now calling the financial shots.
Will that be enough? Harris thinks so, but traditionally companies that undergo recapitalization don't do well. Time will tell if Lineo's fast growth and multiple acquisitions simply made them more vulnerable to one more financial storm or has given them roots deep enough to weather 2002's gloomy economy. One thing is for certain: with hard promises of profitability by mid-year to Canopy and its other financial backers, we'll quickly know what Lineo's fate is to be.
LynuxWorks might be doing the best of the lot with many solid, long term business relationships. It also doesn't hurt any, that its products are strong and that this spring they announced Linux binary compatibility for their LynxOS real-time operating system (RTOS). What may be the key differentiation between them and the other players, though, is their well-thought-out support and certification efforts. On the other hand, they have had several rounds of layoffs. The question here will be whether LynuxWorks can hang in there long enough for their revenue streams to start producing.
MontaVista Software, according to Williams, has a lot of capital and a good technical base. But, from where he sits, their smartest move is that “they are moving beyond embedded Linux standard base to carrier-grade Linux.” Their continuing strong relationship with IBM also bodes well for MontaVista's long term success. That said, in the short run, MontaVista appears to have the largest team still devoted to embedded Linux work and the company also appears to be living off its investment money and not their revenue stream. As is so often the case with embedded Linux businesses, the race between increasing their revenue and running out of funding will determine MontaVista's long term future.
As for Red Hat, their embedded products are good and interesting, but are they really seriously interested in being a major player in the embedded Linux space? Many people we spoke with thought that more and more of Red Hat's attention and resources are being spent on high-end, enterprise Linux projects like Red Hat Advanced Server. While Red Hat's offerings may be the most popular distributed embedded Linux tools around, according to VDC's Linux's future in the embedded systems market research study, Red Hat doesn't seem to be pushing as hard in their marketing and partnering efforts as it used to in this space. What's not clear is whether this is because they don't feel they need to, given the strength of the Red Hat Linux brand, or that embedded Linux at Red Hat has now become a lesser priority.
REDSonic is a small player with a good reputation, especially on the Pacific rim. It's there that their hopes lie. In a recent deal with FSMLabs (whose RTLinux software is also popular in Asia), REDSonic agreed to integrate FSMLabs' RTLinux/Pro real-time Linux extension into REDSonic's RED-Builder embedded Linux development and deployment toolkit, which will broaden its appeal to a wider range of customers. If REDSonic continues to focus their marketing efforts on China, Taiwan, and Japan, they're likely to do well.
Can you make a living by focusing on just one platform? If you're RidgeRun and the product line is Texas Instruments dual-core DSP chips, the answer may be yes. On the one hand, they don't need to support multiple architectures; on the other they've banked their future on this one family. As TI's future goes, so will go RidgeRun's. Still, by not trying to be everything to everyone, RidgeRun may have made the safest bet
TimeSys, on the other hand, is busy delivering real-time and embedded Linux support for a rapidly expanding range of processor architectures, including x86, StrongARM, XScale, PowerPC, UltraSPARC, MIPS, and NEC's VR Series. The real question here is whether the new management team, brought onboard in December 2001, will be able to take all these efforts and forge them into profitability. While an influx of $15.5 million in equity money in March helps, one can't help but notice that TimeSys keeps issuing announcements of new ports to new chip families, but not announcements of new customer deals. Still, with Sun holding some equity in the firm (hence, the UltraSPARC port), TimeSys may well get enough time to see if its broad technology ventures will pan out.
Who will win and lose?
But, in the long run, who will win and lose here? Frankly, we can't tell. LynuxWorks and Red Hat, presuming Red Hat wants it, seem to me to be the most likely to still be in the embedded business in 2005. MontaVista also seems headed in that direction but don't be surprised if in 2005 they're thought of first as a carrier-grade Linux provider and then an embedded player.
On the other hand, I could be dead wrong. The margins are razor-thin, competition is tight, and frankly in the embedded Linux horse race, these horses are too close to call and we're not even close to the final stretch.
not that of LinuxDevices.com or its management.
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.