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

Article: CEO Interview: D. Jeff Dionne of Arcturus (and uClinux pioneer)

Mar 4, 2004 — by LinuxDevices Staff — from the LinuxDevices Archive — 49 views

Foreword: This interview with D. Jeff Dionne, CEO of Arcturus, is the fourth in LinuxDevices.com's CEO/CTO interview series. Arcturus was spun out from Lineo, after having been acquired as RT-Control, which developed uCsimm and uClinux. Enjoy . . . !


Q1: At one point, Arcturus described itself as “the creator and maintainer of uClinux.” Is that still true? What is the relationship between Arcturus and uClinux?

A1: At any given time we have several engineers working on core development and porting uClinux. Our business has taken a much more product-oriented focus, pushing us up the stack and allowing us to leverage our extensive kernel development background, device driver expertise, and innovative hardware designs. We've never found porting and PS a particularly attractive business; but combined with the stable revenues from our controls business, it has allowed us to develop our software stacks and expand our uClinux based products. It has also helped us to develop strong relationships with a number of semiconductor vendors.

The kernel and distribution on a day to day basis is maintained by the SnapGear guys. We host the uClinux.org domains. and work with companies such as Analog Devices, Conexant, or Sigma Designs to help manage their releases.

Internally, we of course use a derivative of the uClinux userland, on MMUless and MMU processors.

Q2: What's happening in the world of uClinux these days? Has being part of the main kernel source tree brought more exposure and uptake to uClinux?

A2: uClinux, by sheer volume of deployments is easily the most widely used embedded Linux variant. Having uClinux in the mainline tree has been the objective since inception, and I'm hoping that with its inclusion in 2.6 the (remaining) embedded vendors will start to offer it in their main products. uClinux is not really different from any other port family.

Q3: Is Arcturus still focused on the Residential Gateway (RG) device market? What are your current projects, and can you mention any clients or manufacturing partners? How have things been since the split from Lineo in 2002?

A3: Arcturus has continually grown since we spun out of Lineo to build a stack in gateways and VPNs. We accomplished this, and now we offer a unique product that scales from RG to enterprise gateways and beyond. We've structured our stack around a very cool MIB (Management Information Base) database engine called MIBflex that relies on simple plugins to enable administration (web, SNMP, UPNP, etc.), an internal XML interpreter, and even an Interactive Voice Response system (IVR) for building telephone decision trees. MIBflex was designed to be a powerful tool for the management of devices and services. In addition to being a solution long awaited by carriers and OEMs, MIBflex also enables us to future proof our products by offering a framework in which we can plug new innovations such as additional media applications.

From major OEMs, early next quarter you will see our SIPstream and MIBflex products in voice gateways running Linux and uClinux. On the shelves, you will find a standalone RADIUS server for SMB, and a wireless VPN product using our AIRmarshal firmware. Conexant, Micrel, Samsung are or will be shortly shipping our BSPs with demo RG software, and our industrial controls products are deployed in the tens of thousands in dental imaging equipment, military communication applications, and non-life critical medical equipment.

Q4: How would you characterize the current state of embedded Linux, as a market and as a technology? Of uClinux?

A4: uClinux is the natural starting point for just about any connected device you would want to build. That said, uClinux (the distribution) has grown to the point that it needs some tools to make it easier to deploy. We hope that the commercial vendors bring uClinux into their product mix, but time permitting, look for some Open Source tools from us on that front. . . .

Q5: What challenges do embedded Linux OS, tools, and services vendors face?

A5: Embedded Linux isn't a mystery. ODMs and OEMs with engineering expertise are perfectly capable of getting Linux onto devices themselves.

You can get a Linux based OS on your board in a day or two provided you choose an SoC with that in mind. That said, the OS vendors face the “roll your own” crowd, and with them it's an issue of providing the tools to get there faster. The tools companies have to help their customer come to the conclusion that it's not in their best interest to waste time porting an OS. Ultimately, for both tools and OS vendors, the message is: product engineers should spend their time building product, not getting the kernel to boot.

Q6: What market opportunities do you see for Linux in the embedded devices and systems market?

A6: Since the beginning we've held the belief that embedded Linux will be everywhere and you'll never even know it's there. I think it's safe to say we've reached that point. Specifically, uClinux is in scores of consumer electronics, including DVD players [see this story –ed], meaning of course that Grandma uses it. (Now, if we could only get her VCR to stop blinking 12:00!) This is significant; it means in the product adoption cycle, we've reached a point of acceptance, and as devices continue to become more feature rich, the competing OSes such as Nucleus will simply fall by the wayside. In digital home devices, our core focus, we've worked with a number of OEM/ODMs who have abandoned efforts in favor of a Linux based solutions.

Q7: What challenges does embedded Linux face as a technology?

A7: A few years ago, I was talking with Kenneth Albanowski and he came to the realization that “Linux is what comes after C”. This is really how we look at it now. Linux and uClinux are “just” the C runtime. As a technology, embedded Linux is hugely portable and successful, from straightforward ports such as Xscale to complex ones like Blackfin.

As a technology, it's a breadth and depth challenge. Embedded Linux must expand, meaning more ports. Simultaneously, the kernel and userspace-level code needs to grow to meet the demands of next-generation products. There is also a third dimension: density. The footprint must stay small, or even shrink. The only way to accomplish all three is through a strong and vibrant community, advancing the base for everyone, constantly tracking chip offerings and deploying widely in real products. What's the biggest threat? Stagnation due to people hoarding code. . . GPL violations in short.

Q8: What embedded Linux technology developments do you find exciting?

A8: For me, embedded Linux was always about building something. It has limited life or purpose on it's own. It's exciting to see a new architecture or chip running (uC)linux, and the second a new box arrives in our office, you bet I'm the first to start ripping into the packaging (much to the chagrin of my co-workers).

Q9: Can you share one or two of your company's most exciting successes?

A9: Without a doubt, it's currently Voice over IP — specifically related to our core focus on consumer products. Consumer VoIP promises to provide the first real widespread “convergence devices.” And Linux, assisted by various open standards and robust implementations, is clearly going to be the engine that will power the next generation of human communications.

Q10: What's your vision of the future for embedded Linux? That is, how big will the embedded Linux market get, and when will it start to reach that peak?

A10: Embedded Linux will be everywhere as a part of the C runtime. The market is another matter. For us, we frequently do the ports ourselves, but when sensible, we've contracted ports. It's interesting to note the number of ports that now come by way of the silicon vendors, submitted to us or other maintainers to put into the CVS. These vendors recognize the value in offering something that helps to develop a product — tools, application software, training, services — for free, both as in beer and speech, and that's what makes it so successful. So to clarify, I think this means there really is no embedded Linux market per se, but rather a market to serve people who are developing products with embedded Linux.

Q11: Any comments on the SCO mess?

A11: When Caldera started down the SCO/Unix path, we thought it was a “Hail Mary” pass for Caldera. Remember, we were part of Lineo then, a sister company. There was no hint of this happening in the future. Like everyone else, when we first heard that about the alleged UNIX rights, we could see what was on the horizon. The folks at Groklaw.net do a great job of sifting through the ever-changing claims and positions in true Open/Free spirit.

Q12: How do you see the 2.6 kernel affecting the embedded Linux market?

A12: Obviously 2.6 represents significant innovation on the part of the community, but remember: use the best tool for the job. The 2.0 kernel is ideal for some embedded products, which is in contrast to the server or desktop market where it's wholly obsolete. 2.6 will expand the suitability of embedded Linux to a larger class of devices, but I don't think the move to 2.6 will suddenly position embedded Linux as the “killer” OS. . . it already was (quite literally for the proprietary OS competition).


About the interviewee: D. Jeff Dionne has over 15 years of experience in electrical engineering, hardware design, and software. In 1995, Jeff became a founding partner at The Silver Hammer Group, developers of test sets and software for high voltage power applications. While working on embedded applications, Jeff pioneered the development of uClinux by applying the Linux OS to low cost MMU-less processors. In 1998, Jeff went on to co-found Rt-Control where he assumed the position of CTO and continued to promote the uClinux OS and supporting software and hardware. In May 2000, when Rt-Control merged with embedded Linux vendor Lineo, Jeff worked in the capacity of Vice President of research until early 2001 when he accepted the position of CTO. Jeff is recognized as the founder of uClinux and has published and presented numerous papers on embedded network systems. Today, as CEO and Chief Architect of Arcturus Networks, Jeff continues to act as innovator, pushing the company forward into the consumer VoIP marketplace with industry leading solutions for the networked home.


 
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.



Comments are closed.