[Discuss] microkernels

[bctill]

Hi all --

I've been thinking more about the problems introduced by code size, and 
how surprising it is to find that this is actually worse for Linux than 
for Windows.

I see that there are a fair number of microkernel projects out there, 
but none of them have any real traction.  I can't figure out why.

According to the "bro science" on the web, the reason is that 
microkernels have poor performance compared to monolithic kernels.  But 
that is obviously not true.  Just look at the Windows kernel compared to 
the Linux kernel.  Or, if you want a truly stellar example, look at QNX.

For a consumer desktop, you need extraordinarily wide-ranging hardware 
compatibility, for enterprise servers, the range of hardware that must 
be supported is quite a bit narrower but still pretty broad.

But I can think of no reason why people don't use mircokernel-based 
operating systems for small-scale, mission-critical applications, for 
which reliability and security are crucial.

Why does no one run servers on QNX, for example?

How about Debian GNU Hurd?  Can you run a server on that?

The Mach and L4 microkernels are out there, but the projects that 
coalesce around them keep dying.  Why?

Microkernel architechture seems like the obviously optimal platform that 
FOSS should move toward.

Why is there so little interest?

Regards,
-- Bernie.