David Mosberger on Linux, Penguins and Moonwalks

Biography and interview transcript by Jamie Beckett


David Mosberger is Linux central at HP Labs. Although he doesn't decorate his cube with the signature Linux penguin logo, he is a long-time user and a contributor to the open-source operating system. Currently, he's working furiously to make sure Linux works on the new IA-64 processor Intel is readying for release next year. The architecture for the IA-64, Intel's first in a series of 64-bit processors, is based on architecture developed jointly with HP.

Mosberger grew up in the small town of Niederweningen near Zurich. He holds a professional degree as an Electronics Engineer, an HTL Diploma (BS) in Computer Science from HTL Brugg-Windisch, Switzerland, and Master's and doctoral degrees in Computer Science from University of Arizona. He is a voting member of the Association for Computing Machinery (ACM), the IEEE Computer Society and USENIX.

Since September 1997, he's been a member of the technical staff in HP Research Labs. His research interests are in operating systems, high-performance and scalable Internet systems, computer architecture and compilers. Before starting the Linux work, he worked on improving the quality of service for Internet services.

Q: How did you get involved with Linux?

A: Linus (Torvalds, the inventor of Linux) started to work in Linux in 1991. At that point, it was just for the PC platform. I was a student at the University of Arizona pursuing a Ph.D. in operating systems, and we were working on our own experimental operating system based on Digital's Alpha chip. I started to realize it takes a lot of time to maintain the low-level aspects of an operating system such as device drivers. In other words, just maintaining the basics of an OS takes a lot of time. So I thought that if Linux ran on the Alpha, there would be an entire open-source community to maintain and enhance it and since Linux is open-source, I could borrow the low-level parts for our OS. That would allow me to focus more on the parts of the work that are interesting from a research perspective.

Q: Since February 1998, you've been laboring to make sure Linux works on the IA-64. Why is it so important to make sure this happens?

A: Linux at HP is at about the same stage that Windows NT was a couple of years ago. But if Linux becomes market force like NT, HP could not afford not to have offerings in that space. When I first started work on this in 1998 it was quite speculative -- hardly anyone in a suit knew how to pronounce or spell Linux at the time.

Q: What do you like about Linux?

A: The open-source nature of Linux makes technology transfer easier. You can share things more easily. It's a relatively simple, easy-to-grasp operating system. You can build a new kernel (the central module of an operating system) in about five minutes. To do the same with Unix or NT takes an hour or so. So it's easier to do something new on Linux. Another thing that's nice is that Linux tends to run very stable, so it's nice as an Internet server and even as a desktop environment.

Q: Is Linux leading a trend toward open-source software?

A: It certainly is a trendsetter. I think Linux is far from having peaked. Exactly how far it will go is uncertain. Microsoft is powerful. But there are viable business models for open source -- Red Hat's an example. The Linux distributor showed you can make millions of dollars giving stuff away. (The company, which recently went public, has a market capitalization of more than $7 billion.) I think Red Hat is the first in a series of companies that will make money in open source. The question of whether open-source will go all the way to applications is still open, though.

Q: How do you feel about penguins?

A: I'm not sure I've ever met one. I like the penguin as a symbol for Linux. It's so non-corporate, this fat, smiling penguin.

Q: How did you wind up working in technology, and particularly on the software side of engineering?

A: As a boy, I was very good at taking things apart, but not so good at making them work. So I wasn't so adept at the hardware side of things. As a teenager, I was always interested in technical things. Somewhere along the way, I saw that calculators could handle floating-point numbers even though they were using a binary system. I didn't understand how it worked, so I kept reading until I figured it out. What I like about computers and software is that the only boundary is your imagination. I don't look at it as an engineering discipline. I look at it as an art form. I enjoy coming up with better ways to do something. It's an infinite space of challenges and solutions that are limited only by your thinking, not by physical boundaries.

Q: What else were important moments when you were growing up?

A: A defining moment for me was the moon landing. I was three years old. I probably didn't understand half of it, but I stayed glued to the TV. In Switzerland, the moon landing happened late at night -- I think around 11 p.m. -- in other words, way past the bedtime of a 3-year-old.

Q: How did you come to establish one of the world's first WebCams?

A: At graduate school, you get these cubes with no windows. It's a really dark place, and they crank the air conditioning up so it's freezing. I thought at least I could have a virtual window so I could see what was going on outside. That was in the mid 1990s, and it's still posting pictures of the skyline over the university (in Tucson) and people are still looking at it, although the picture quality of the camera has, unfortunately, deteriorated a lot since it was installed. If it breaks, they (at the university) get mail within five minutes.

Q: What's ahead for you?

A: Who knows? Lots of excitement. I expect we'll be able to transfer the Linux/IA-64 technology to various HP divisions over the next couple of months. After that, I'd like to get back to working on Web quality of service (QoS) issues. The idea is to use Linux, and the Internet itself, to enhance end-to-end Web quality of service. Right now, Web QoS is focused on the server side, but it doesn't do anything on the client side or anywhere in between.

Q: Are you ever tempted by startup fever?

A: At this time, I'm more involved in business development than research, so it feels very much like a startup. Except I don't have to deal with the capital side of things and can focus on the technical aspects. Selling my soul to venture capitalists is not on my list of things I want to do.

Q: Do you miss Switzerland?

A: It would be nice to have the Alps close. But I don't miss the weather for a second. I love working in the Bay Area. The opportunity to make things happen is so unique here.

Copyright 2002