Messages in this thread | | | Date | Sun, 30 May 1999 21:42:33 -0700 | From | Oscar Levi <> | Subject | Re: is Linux obsolete? |
| |
On Sun, May 30, 1999 at 10:39:24PM +0100, Alan Cox wrote: > > The quintessential microkernel operating system is Mach. It is based > > on passing messages between independent services. The memory > > Please don't use Mach as an example. Mach is a research tool, it is very > big, it nowdays uses the same address space for the kernel too.
Oh no. I'm not a Mach fan.
Since the original question was about the nature of microkernels, Mach is a well-known example from a design standpoint. Amiga's success is hard to measure since the product failed commercially, albeit in part by Microsoft's shady maneuverings. Mach lives in the NeXT. Another commercial product using message-passing microkernel is called Chorus, produced by a French concern. And, there's QNX as you mentioned. NT would not be characterized as a message passing system, nor as a microkernel by any discerning designer.
As far as my studies went, there's a general belief that message passing tends to be a little slower than kernels like LT's. A now-defunct project called SPIN at the University of Washington developed a traditional, procedural kernel using a modified version of Modula3. They enforced protection within the kernel by using this strict type-safe language. The idea worked out OK. You could inject a Web server into the kernel and see impressive throughput. But then again, you had to code in Modula3.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |