MontaVista weighs in on Intel’s Wind River play
Jun 5, 2009 — by Eric Brown — from the LinuxDevices Archive — 1 viewsWith the news of Intel's impending acquisition of Wind River still reverberating, we asked rival Linux distribution and tools firm MontaVista Software to offer its perspective. In an interview, MontaVista VP of Marketing Joerg Bertholdt (pictured) sounds off on MontaVista's future Intel support, and much more.
Since Intel announced it had signed a definitive agreement to acquire Wind River for $884 million yesterday, the web has been ripe with speculation about the impact of the impending deal. MontaVista Linux is one of the top two embedded Linux distribution and development environments along with Wind River Linux, so we decided to check in with MontaVista's Bertholdt to get his company's views (see farther below for the interview).
Both MontaVista and Wind River are based in Silicon Valley and offer a similar mix of Linux offerings, including Linux kernel distributions, middleware, embedded design tools, and services. They both serve a broad array of product types and industries, including Carrier Grade Linux (CGL) networking products, industrial equipment, consumer electronics, and hot new mobile devices such as smartphones.
Yet, there are also key differences. Wind River is public, while MontaVista is private. Wind River is the larger company, largely due to its proprietary VxWorks real-time embedded business, which still dominates revenues despite the fast growth of its Linux business. Compared to MontaVista, which was a pioneer in the embedded Linux business and is still dedicated solely to Linux, Wind River moved relatively late into embedded Linux.
MontaVista has until recently dominated mobile Linux, with a longstanding partnership with Motorola, among others. Yet, Wind River has come on strong in that fast-growing market, and has become particularly influential in the Linux Mobile (LiMo) Foundation, as well as with Android.
A disruption in the force
It seems clear that embedded Linux has been improved and empowered in recent years by the dynamic rivalry between the two firms. Although the acquisition suggests that embedded Linux will only grow in power and influence in the technology world, the lopsided nature of a market dominated by an Intel-owned Wind River will almost certainly disrupt that dynamic. While cheered by the seal of approval that the acquisition brings, many Linux developers have a queasy feeling about the ramifications of an industry that shows signs of being dominated by giants like Intel, Microsoft, and Google.
Then again, the same fears were raised when IBM got the Linux religion with servers, and the Linux server market has thrived ever since. Meanwhile, the old ways are already changing with the emergence of increasingly sophisticated semiconductor-specific development platforms, as well as new embedded Linux development platforms from Embedded Alley, Timesys, and others. The “traditional” development model is being stretched, with the once-clear lines breaking down between DIY and commercial approaches. MontaVista itself has just radically transformed its MontaVista Linux with a new version 6 that breaks down the product by semiconductor architecture in so-called “Market Specific Distributions” (MSDs).
Android vs. Moblin
Meanwhile, in mobile consumer electronics, open source projects like Android and Moblin offer fast-growing platforms of maturing tools that compete at least in part with the commercial embedded Linux development market. Much of the speculation in the press about the impending buy-out has focused on these platforms, although in truth, netbooks, smartphones and the like are still a relatively small part of the embedded Linux universe served by Wind River and MontaVista.
Still, mobile Linux is where the action is, as evidenced by the avalanche of Android and Moblin announcements this week at Computex. Some have suggested that the acquisition is a shot across the bow of Google to inform the search giant that it cannot so easily waltz in and dominate the embedded market as it has done with, well, just about everything else it's touched. Others say it's also a warning shot to Apple and its insanely great iPhone.
Under Intel, Wind River is likely to start distancing itself from the Google-sponsored Android. Although job one will be even more tightly aligning itself with the netbook- and MID-oriented Moblin, Wind River is also likely to continue to support the LiMo platform, where it is too deeply involved to easily pull out, even if Intel wanted it to do so.
There are signs that a future version of Moblin will support telephony, bringing it directly in competition with Android. Intel and Moblin are collaborating with Nokia and its Maemo project on an oFono project that could bring telephony support to both these Linux-based open source projects. The project opens up the potential of a Nokia/Intel axis that could take on Android together. It may be far-fetched, but even the Soviets and the Americans were able to come together in the pursuit of winning World War II.
Speaking of open conflict and failed alliances, the punditry seems to agree on one thing about the acquisition: that this is the final Intel insult to former soul brother Microsoft. Some have even suggested that Wintel be replaced with “Lintel.”
A conversation with MontaVista's Joerg Bertholdt
To get MontaVista's view of the Wind River acquisition, we asked Joerg Bertholdt, VP of Marketing, for his thoughts. The following offers excerpts from the conversation about the buy-out, Moblin, Android, embedded Linux, and more:
LinuxDevices: “With the pending acquisition of Wind River, Intel suddenly becomes a MontaVista rival in addition to a being a partner. Does this change MontaVista's support for Intel processors such as the Atom?”
Bertholdt: “It does not change our support. We have a significant number of customers on Intel architecture, and we will continue to support them while we also position our customers across all the major semiconductor solutions. Our assessment is that Intel isn't really trying take out commercial embedded Linux support from the ecosystem. We certainly see parallels with Intel's approach to other products. For example, Intel already sells Intel optimized compilers, yet it also enables a broad ecosystem of third-party compilers that target Intel architecture. We believe we still have access to the right level of detail to Intel technical information and roadmaps to build products of value to our customers and to not be at a competitive disadvantage.”
LinuxDevices: “MontaVista recently joined Moblin and then announced a Market Specific Distribution (MSD) for MontaVista Linux 6 that supports the Intel-sponsored Moblin v2 stack running on the Intel Atom. Does your open-ended position in regard to Intel chips also extend to your support for Moblin?”
Bertholdt: “We will continue to support Moblin. What we are hearing from the Moblin team is that Intel still wants a broad ecosystem around Moblin and the Atom. Intel really went strongly on the open source route with Moblin, and has handed it over to the Linux Foundation. We doubt that that Intel is going to backtrack and pull things back in. Because Moblin is an open source project, it really puts us into the position of adding our value around it.”
LinuxDevices: “Intel has stated that it expects Wind River will continue its ARM and PowerPC support after the acquisition. Do you think this will happen?”
Bertholdt: “I would say that Intel's statements that it will be 'business as usual' at Wind River represents the typical positioning you often find in pending acquisitions. But typically in these cases, it doesn't make business sense in the long term, and the positioning often changes quickly. People are taking it with a grain of salt, I think. We certainly expect that Wind River will take an 'Intel first' approach to its hardware support. The question is whether it will be 'Intel optimized' or 'Intel only.' Certainly, Intel architectures will be ahead in terms of support compared to other architectures.”
LinuxDevices: “If Wind River does end up overly favoring Intel architectures, does this offer an opportunity to MontaVista and other embedded Linux tools and distribution competitors?”
Bertholdt: “It is certainly going to be a time of uncertainty in the industry, but it is an opportunity for us. We have already received numerous calls from concerned Wind River customers working on non-Intel architecture projects who are worried about continuing support. We are now in an even stronger position to provide our customers a choice of architectures. We see it as a validation of our MontaVista Linux 6 approach. Part of it will depend on the future support structure that Wind River and intel will provide for non-Intel architectures. If we start to see a significant time lag between the amount of time it takes Wind River to support new Intel silicon versus the time it takes to support new processors from other vendors, it will make it more difficult for their customers. We will just have to see how the support infrastructure rolls out, and how closely Wind River aligns itself to the Intel hardware.”
LinuxDevices: “On the other hand, aren't you concerned that you are suddenly in competition with a giant like Intel, with all of its marketing clout?”
Bertholdt: “Intel is first and foremost interested in selling silicon, so it remains to be seen how much from a marketing and sales perspective they will really pour into selling Wind River software. If you look at Wind River's financials, a lot of their funding of Linux was siphoned off from profits earned from VxWorks. By comparison, we have been providing support for embedded Linux for over a decade now. We offer stability.”
LinuxDevices: “Stepping back from the MontaVista perspective, what do you think the acquisition means for embedded Linux? Is it a good thing to have a technology giant like Intel controlling an increasing share of embedded Linux, first with the growing clout of Moblin, and now by owning a major tools and distribution vendor?”
Bertholdt: “We see the acquisition as a validation of the embedded market in general. It's an often misunderstood market, an odd market with all its diversity, and if the world's largest semiconductor is saying that the embedded market is important to them, then it's good for the embedded market and the embedded Linux market, and a confirmation for MontaVista.”
LinuxDevices: “Do you expect other technology giants to follow suit and invest in the embedded Linux development market and acquire a company like MontaVista?”
Bertholdt: “We cannot speculate on that.”
LinuxDevices: “Aside from the benefits of the acquisition in pure business terms, is there a big advantage to Intel having its own Linux tools business in-house? Some have speculated that Intel might use Wind River's compilers to help it optimize software for its low-power x86 chips.”
Bertholdt: “I doubt Intel would acquire a company for close to a billion dollars just to get their hands on a compiler.”
LinuxDevices: “Some have suggested the acquisition is not only a distancing of Intel from Microsoft, but, along with the big Moblin push, a preemptive move against Google to keep it from dominating the mobile device business with Android. Is this a chess move against Google, and how directly do you see Moblin and Android competing?”
Bertholdt: “There's certainly some possible overlap between Moblin and Android, especially for devices driven by an application ecosystem of independent applications. There is a segment that is a bit of a gray area between enterprise desktop computing and embedded computing.”
LinuxDevices: “With Android being ported to MIPS for a broad array of consumer electronics and industrial equipment, do you see Android, or even Moblin, becoming defacto stacks for embedded Linux devices in general?”
Bertholdt: “I don't think so. The embedded market is very broad, and the majority of devices are driven by independent applications. You are probably not looking for cool apps to add to your medical device or your network equipment. There is still a gap between the consumer app-generation market and the broad embedded market where things like real-time capability or configurability are more important. If you go back to the '90s, and the promise and hype of Java, well, it's an important technology, but it didn't quite happen the way it was projected. There is a common misperception about devices that are driven by applications and those that are more dedicated-function devices. It's important to really understand the requirements of different market segments. Still, there are places in the embedded world where parts of Android and Moblin can be leveraged and dropped in.”
LinuxDevices: “With the growing split in the Wintel alliance and Intel's gamble on Wind River, do you see Microsoft being inspired to reinvest their efforts in the embedded market with Windows Mobile and Windows CE?”
Bertholdt: “I cannot speculate on that, but Linux is the defacto standard of the embedded market. That train has left the station.”
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.