DragonFly BSD
DragonFly users List (threaded) for 2004-10
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: Our fearless leader on video!


From: Steve Jothen <sjothen@xxxxxxx>
Date: Wed, 27 Oct 2004 13:34:05 -0700

Added, changed some stuff and came up with this:

I should talk about DragonFly, and I guess a little more background is
required before I talk about DragonFly. FreeBSD was not the first
Operating System I worked on. I actually worked quite a bit on Linux
before FreeBSD, and even during my Linux days, I also wrote a lot of
embedded Operating Systems for work that I was doing up in Tahoe. We did
Telemetry systems, so, we had Linux as a base station and later a NeXT
machine as a base station, and finally now we have a FreeBSD machine as
a base station. But the units out in the field were all these embedded
Operating Systems, and of course to interface
with Linux, NeXT, or FreeBSD, we had to use a standard protocol, so it
talks UDP. So the Concept of working on an Operating System is something
thats really been embedded in my life for a very long time. My work on
FreeBSD 4 was a wonderful experience despite having to pull to get
things in, push and push and push. But with FreeBSD 5, they kind of went
off in a direction that wasnt really compatible with my own roots and my
own desire in terms of future direction of FreeBSD. In the embedded
world, the Operating Systems I wrote were very heavily threaded, and
they were very heavily event oriented and messaged. I knew that this
type of Operating System could be written to operate very efficiently, I
mean these embedded systems were running 10MHz 68000's and they were
doing a lot of stuff. So that division is what caused me to start
contemplating doing DragonFly. And I didn't actually announce it, I
started working on the Light Weight Kernel Thread Scheduler almost
immediately in January last year and I didnt announce DragonFly until
June last year. I spent all that time building all the LWKT stuff from a
4.X base and making sure that it did what I wanted, that it worked, and
that the project could really be announced and go forward. And infact
thats what happened, I threaded the interrupts, I got LWKT working and
announced DragonFly and got enough interest that I felt that it could
move forward as a project. From then on its just been great, I've been
pretty much overloaded. I can still do programming, I am not 100%
manager like Linus Torvalds is thought to be. Its been a great ride, and
its going to continue to be a great ride, we have got lots of very
interesting things planned for DragonFly. Our ultimate goal is SSI, but
to get there, we have to go along a parallel track with what FreeBSD 5
is doing for quite a while, but we will start to diverge. The basic
things like getting rid of the big GIANT lock, our approach to that is
very different from FreeBSD 5. Once we get past that it gets far more
interesting, it gets into Virtualized Machines, Virtualized Resources,
Cache Coherency across clusters, so it will be interesting.




[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]