News Archive (1999-2012) | 2013-current at LinuxGizmos | Current Tech News Portal |    About   

Ten years after: An Interview with MontaVista’s Jim Ready

Jul 24, 2009 — by Eric Brown — from the LinuxDevices Archive — 6 views

Embedded Linux pioneer and MontaVista CTO and founder Jim Ready recently spoke with LinuxDevices about how the industry has changed over the last decade. The interview arrives on the heels of MontaVista's 10th anniversary.

LINUXDEVICES: In 1999 when you founded MontaVista, desktop Linux was in its infancy, the boom in Linux servers was still several years away, and the embedded world was devoted to real-time operating systems. What made you think Linux would be accepted or even work in the embedded world?

READY: No analysts were covering Linux at all back then. Red Hat was well known, but I think that like a third of the revenue came from selling T-shirts and stuff at Fry's. Linux was considered to be very wacko and was derided as crazy. Bringing Linux to embedded was even crazier, people told me. It was way too big, way too slow, it was not real-time, and it was written by 16-year olds and Communists. But at Ready Systems, we were familiar with VRTX and the RTOS world, and customers kept asking for the RTOS kernel to have memory protection and to be POSIX-compliant. Essentially, they were asking us to make a real-time UNIX.

LINUXDEVICES: But not, probably, a real-time Linux.

READY: I had been following Linux, seeing its steady improvement, and I was getting fascinated by open source process. I saw Linux getting ported to other architectures. But the thing that pushed me over the edge was an article in Linux Journal about a radar controller. The first version had two multi-bus boards, one running UNIX and the other running VRTX, but then in the second version, the UNIX side was replaced by Linux. It became clear to me that in the highly fragmented embedded OS industry, which was really a stepchild to computing back then, Linux, which had the memory bus support and richness of connectivity, could have a revolutionary effect on embedded computing. First, of course, we would need to enhance it and fix the technical limitations such as size, and availability, and move toward real-time, but I knew we could do it.

LINUXDEVICES: You must have met some resistance in those early days.

READY: When we announced [Hard Hat Linux], the RTOS guys said it was a joke. But over time it took hold because it was really driven by the applications that customers wanted to build. The device makers wanted more sophisticated, connected, and robust devices, and they wanted a standard, so eventually it became impossible for the RTOS industry to compete. Linux turned out to be the center of gravity.

LINUXDEVICES: Several years later, MontaVista was instrumental in moving Linux into cellphones. It was considered quite risky at the time, but now, of course, Linux is growing fast in the smartphone market. How did you manage to get the mobile Linux industry moving?

READY: You have to give credit to Motorola and NTT DoCoMo for that. Back in 2003, we knew we could do it, but we weren't proselytizing about it. When we would mention it, people would say, "Oh yeah, you're going to put a server OS on phone. Right." But then we discovered that Motorola and NTT DoCoMo thought it possible, too, and we got together to do it. That was a bold move on their part, real out of the box thinking, especially considering that it was in the middle of the SCO lawsuit. But we knew that the ARM processor was capable, and that we could get a nice fast boot. It happened faster than I thought it would.

LINUXDEVICES: Was your eventual success with embedded Linux due in part to the fact that the embedded industry was, and continues to be, so fragmented? Desktop Linux continues to struggle in an environment dominated by Microsoft, but although Microsoft is a major player in embedded, they're far from dominant.

READY: Yes, the landscape was very fragmented, but there became a realization of the threat that Microsoft represented if they were to get their hands on consumer devices. Microsoft tried, but Sony said we won't let Microsoft tell us how to build a television. The embedded world wanted to deny Microsoft another monopoly, and they realized that to succeed at that, they needed to be less fragmented. They needed a standard. On the desktop, Microsoft had the tyranny of the apps, which is not really an issue with most of the embedded market. In areas where apps are an issue, such as cellphones, the industry has actively organized to not buy into Microsoft.

LINUXDEVICES: Does the open source nature of Linux help drive that resistance to dominance by a single company?

READY: On some levels. The brain-dead lock-in can't occur with open source, where a company can hide away their source code, which is significant. There are other forms of lock-in though. For example, Wind River can easily move its VxWorks customers to Wind River Linux, but they don't have the same kind of lock-in when it comes to new projects. Open source also enables consortiums like LiMo to operate, so companies can band together and not duplicate efforts.

LINUXDEVICES: You mentioned that the fast growth of mobile Linux surprised you. What other surprises have you seen more recently?

READY: One thing we underestimated was the number of semiconductor engineers and partners working on Linux to enable Linux on chips. That happened faster and at a larger level than we thought. When we started out building MontaVista Linux earlier in the decade, we had to make the assumption that Linux may not exist on a given piece of hardware, or that if it was available, the kernel version might be four or five rev's back. Beyond that, Linux was still somewhat unstable. So we developed a set of requirements and industrial processes to produce a commercial quality Linux that could go into any device. But now Linux has gotten really deep into the fabric of the chip business, and we are now adjusting to that reality with our new MDE approach in MontaVista Linux 6.

LINUXDEVICES: Don't things get more complicated trying to integrate all this with more open source software and stacks like Android and Moblin? Isn't that reality of growing complexity also reflected in MontaVista Linux 6, with the new integration engine and content server? Is there more of a focus on process than on fundamentals?

READY: Developers are integrating more and more open source code from multiple sources, and all these pieces are both interdependent and independent. That's breaking all the rules of software engineering. All these components are developed independently of one another and they change all the time, and then there are always some other subsystems that were not built by the same group, so it ends up breaking. The open source process is vibrant and instrumental, but it has these bad properties in some sense. The larger good of open source is worth it, but you can't underestimate what you're getting into. If you're drawing on sixty million lines of constantly changing code, it's not going to be easy to deal with. So some of these realities are behind MontaVista Linux 6 — we're helping developers integrate all these pieces.

LINUXDEVICES: What about multi-core? Doesn't that make these development projects even more insanely complex?

READY: Multi-core is a fundamental challenge, but there is no single solution. Virtualizaton is one solution, as are analytic tools and simulation tools, but no one vector like virtualization will overwhelmingly be a slam dunk.

LINUXDEVICES: With Google backing Android and now the new Chrome OS, and Intel acquiring Wind River, Linux firms are now in the company of some pretty big players. What are the opportunities and dangers of such involvement?

READY: In general, the industry's collective consciousness is sensitive to anyone getting a Microsoft-type hold on things, and open source mitigates that. Nothing is going to sit across everything, and that's good — it makes for a more vibrant industry. We have experience in the Android side, and we may see Chrome OS move down into the embedded space in two years. Meanwhile, what we've done with Dell with Latitude ON and MontaBello is moving toward the idea of a cloud computing operating system. What Chrome OS is supposed to do a year from now, we're already doing with MontaBello today.


This article was originally published on and has been donated to the open source community by QuinStreet Inc. Please visit for up-to-date news and articles about Linux and open source.

Comments are closed.