To view parent comment, click here.
To read all comments associated with this story, please click here.
Well, off the top of my head, how about this:
http://lwn.net/Articles/220255/
Who tried to "go" microkernel? :-) Do you know anything about QNX design? IIRC, their OS was microkernel from the very beginning. In fact, the kernel was very minimal and everything run in the user space, even drivers.
Amigans surely remember that design, right? IIRC even on Amiga, you could stop and start new driver, eventually without reboot. And that is the point I met with QNX - during 1998, Amiga under the wings of GW, was supposed to switch to new kernel. And QNX was chosen. I met Dan Dodge from QNX in Koeln, and he is really a smart guy.
Back at that time, when later Jim Collas left Amiga, there was switch suddenly announced - switch to Linux kernel. And in 1998 it was us, Amigans, who in fact were scared, because there were clear figures of responsiveness of those two kernel designs. And I even remember Linus joining QNX usenet just to be educated by some russian guys. QNX was heavily used in nuclear plants, automotive industry, or for applications like blood filtering during the surgery. At that time, you surely would not like to have it running under Windows or Linux, none of which was - realtime.
My kudoz to QNX and what it represents - very clever design.
Well, now back to world, which is overtaken by systems like Windows Vista :-)
Cheers,
Petr







Member since:
2005-07-24
"""
QNX is doing extremely well in its intended space
"""
Granted. Microkernels can do well in niche areas.
It seems that whenever a general purpose OS developer tries to go microkernel, they end up backpedaling and end up with a retrofitted monolithic design.
Except for Minix which, despite what Tanenbaum claims during his combination dog and pony show / microkernel pep rallies, performs abysmally.
And, of course, for The HURD. That shining example of the power of the microkernel design concept.