Subject: Re: microkernels
To: None <tech-kern@netbsd.org>
From: felix zaslavskiy <felix@students.poly.edu>
List: tech-kern
Date: 07/02/2003 16:50:04
This is loaded question. Existance of Linux is partly due to the
micro/mono kernel design difference.  I would take a look at the
Tannenbaum-Torvald Debate (flame?)  Its reprinted in book OpenSources
but its bound to be somewhere online. And yes if you read Tannenbaums
Modern OS book which many college level OS courses use he does still
advocate microkernel design. I would also look into the design of Mach.

I dont see how its relevant to NetBSD in any case.


On Wed, 2003-07-02 at 16:23, Ivo Vachkov wrote:
> Since most of us here are interested in OS development I think it's a 
> good idea someone to introduce pros/cons of MicroKernels vs. Monolithic 
> Kernels OSs :) Further more, QNX is neither OpenSource, nor Free so in 
> one way or another it's hard to get information on it's design. So the 
> point is when you have the time write your "From system developer's 
> point of view: moving from true network distributed micro kernel OS QNX4 
> to solid kernel OS NetBSD" :)
> 
> Ian Zagorskih wrote:
> > On Wednesday 02 July 2003 19:17, Gary Thorpe wrote:
> > 
> >>>Before NetBSD i worked alot with microkernel OS QNX4 and i cannot say
> >>>it's
> >>>supperior or not. Just another ideology. Every architecture has it's
> >>>own
> >>>benefits and weak sides.
> >>
> >>Yes, but some architectures are better than others, despite all having
> >>benefits and weaknesses (of course everyone has differing opinions on
> >>which ones are better). Would you like to replace your modern
> >>multitasking OS with CP/M or DOS 3.3 - after all everything has
> >>strengths and weaknesses? It depends on what you want to do with the
> >>OS. So what are microkernels well suited for as compared to monolithic
> >>systems? In what cases would you opt for one over the other?
> >>
> > 
> > 
> > Ohh well, this is a long topic and i'm really not so long in BSD world, less 
> > then two months. After five or six years in embedded design for QNX4.x. And 
> > this is 2:38 of night atm here so i'm a bit tired to type a long objective 
> > letter :)
> > 
> > But this topic is quite interesting [IMO] and i found something good 
> > developing embedded stuff for NetBSD now same as i lost some good QNX 
> > specific features i used to use in the past. So if you want i can at least 
> > try to write a bit later some esse a'la "From system developer's point of 
> > view: moving from true network distributed micro kernel OS QNX4 to solid 
> > kernel OS NetBSD".
> > 
> 
>