Re: User vs. Kernel (was: To be smug, or not to be smug, that is , the question)

Albert D. Cahalan (acahalan@cs.uml.edu)
Sat, 23 Jan 1999 05:27:14 -0500 (EST)


Jon M. Taylor writes:
> [gone]
>> Jon M Taylor <taylorj@ecs.csus.edu> writes:

>>> Soon, probably within a year or two, a decent, open-source
>>> next generation OS like Hurd (Microkernel), FluxOS (polymorphic
>>> virtual machines/nested processes), Dolphin (exokernel), The
>>> Cache Kernel, or another NGOS will be developed to the point
>>> where it can step in and cannibalize the guts out of Linux to

> Right again. What will happen is that there will continue to be
> a shakeout in the NGOS field, and one or two of the designs will absorb
> the good features of the others and come out on top. If they proive
> clear and present benefits, they will be used as long as porting apps
> is not that difficult.

Most designs are not serious about performance, compatibility,
and real-time. Very few designs are serious about security and
persistent system image.

The HURD people don't seem to care about _any_ of those things.
There is a FreeBSD split that might try for scalable performance
(not low-level uniprocessor performance) and compatibility.
Cygnus is doing real-time and not much else.

There goes the "clear and present benefits".

> Unix has a number of serious defects:
>
> * Lack of generalized message passing
> * Lack of a decent privilege/capability model
> * Blocking I/O
> * Interruptible system calls
> * Process-based instead of interrupt-based

Any UNIX-killer must also have:

* performance like Linux
* compatibility with Linux
* hard real-time support
* security at the B1 or B2 level at least (call it B1.5 if you want)
* persistent system image
* cool name like Linux (serious, these fail: GNU Hurd FreeBSD 386BSD)
* good support for Win32 emulation

I'd like to remind you that security does not exist if it must be
disabled for ease-of-use or Unix98 compatibility.

-
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/