|
|
Executive Interview: Michel Gien and Michel Genard of Jaluna
2005-01-21 Article Rating:  / 0
Foreword -- This Executive Interview includes two top officials at Jaluna, a Linux-oriented virtual platform software company that was France's fourth-best funded startup in 2004.
Jaluna's "OSware" technology -- sometimes called the VMware of embedded -- enables multiple operating system instances to share the same hardware platform. It targets telecom equipment, mobile phones, set-top boxes, and more. Included in the interview are CEO Michel Gien (left) and VP of Marketing and Business Development Michel Genard (right), who departed Wind River to join Jaluna in November of last year. Topics covered include Jaluna's history, funding, technology, design wins, market strategy, and more. Enjoy . . . ! Q1. As we understand it, Jaluna was formed in August 2002, by the creators of Sun's Chorus microkernel operating system (story). Could you provide a brief background on the company and key players? A1, Michel Gien -- Our team has been working together for about 15 years. I was one of the co-founders of Chorus Systems in 1987, and we developed the company around Unix-based operating systems. We developed a microkernel architecture that we worked out with [a company that was called] Unix Systems Labs at the time, and then we focused on telecom equipment manufacturing, both infrastructure and devices. In mobile phones, our Chorus kernel is still in Alcatel mobile phones, for example. In 1997, our team was acquired by Sun Microsystems. At the time, we were about 80 people. We had subsidiaries in the US and Japan, and within Sun we worked on two projects. Initially, we worked on Java devices -- what was called the "Network Computer" at the time -- and on other devices such as webphones. The integration of Java into devices. Then we moved more into more carrier-grade telecom cluster systems, with a combination of real-time kernel and Solaris. We spent five years at Sun, and when Sun wanted to refocus on the computer server market because of an economic downturn, our team was still the original team, and so we proposed to Sun to actually spin us out. Sun was very happy about that. So we started Jaluna in August of 2002, with two directions. The first one, which allowed the company to really get started, was that we took over some of the Sun support and service business around Chorus for some customers that Sun did not have the resources to pursue. That allowed us to get some initial business that helped us support our new technology work, developing what we now call OSware, our virtualization platform. We worked on OSware for a year and a half, almost two years, and in the middle of last year, we got the first round of investments by institutional investors. That's when we got Atlas Ventures, Index Ventures, and Cisco Systems. That's the history. Q2. What was your real-time kernel called? A2, Michel Gien -- ChorusOS. It was a real-time kernel with POSIX interfaces and high scalability [including] a very tiny version [usable] in mobile phones, and high-end versions [usable] in network infrastructure systems. Our main competitor originally was the Mach operating systems, if you remember, back in the mid-80s. Q3. The one used in Apple's original version of Linux? A3, Michel Gien -- Yes. But Mach was very Unix-centric. Chorus could support Unix, but also focused on real-time applications, and providing determinism and real-time response with a wider range of applications. Q4. When did you first start to work with Linux? A4: Michel Gien -- When we left Sun, most of our work on UNIX was around BSD, and we were using a lot of the BSD software for file systems, process management, stuff like that. But it was obvious that irrespective of the technical merits, everyone was moving to Linux. The main reason was the existence of a wide community of developers. Linux was clearly the one getting the most traction. And Linux is Unix, one way or another, so for us it was not a big deal to adapt our technology to Linux. In mid-2002, we started work developing a cross-development system, because we wanted to target embedded systems. We developed a complete distribution for developing Linux on a host machine for a target system. And then we developed what is now our virtualization technology, which we want to be the focus of the company. We are not in the Linux distribution business, and we don't want to just try to improve the Linux kernel itself. There are enough people doing that. We want to leverage Linux, stay compatible, and stay in the Linux tree. But the purpose of Jaluna is to add value on top of Linux-based systems by leveraging the experience and technology we have building operating system solutions. We leverage our experience to build operating system solutions that mix and match different types of operating environments, including real-time and general purpose Linux. Before, we used to build Unix functionality on top of the real-time microkernel, with access to the real-time APIs for the real-time applications, and access to Unix applications from the Unix API. The problem with that approach is that it's quite difficult to follow up with the evolutions of both the real-time aspects and the Unix aspects. When we started Jaluna, we thought that the best way to combine environments like that was to actually keep the full kernel as unmodified as possible, and do the switching between the operating environment at a low level, close to the hardware, rather than stacking up operating system APIs on top of each other. That's basically what we got from our Chorus experience. But in terms of components and functionality, the Jaluna technology is quite different.
|
news feed
