Subject: Re: Hard realtime scheduling
To: SODA Noriyuki <soda@sra.co.jp>
From: Jonathan Stone <jonathan@dsg.stanford.edu>
List: tech-kern
Date: 04/11/2005 16:18:43
In message <200504050720.j357KND25066@srapc342.sra.co.jp>,
SODA Noriyuki writes:

>Not requiring special privilege is problematic, since a realtime
>process can occupy whole CPU resource.
>Is there any prior art about such feature in UNIX-based OSes?

Oh, c'mon. ``The fair-share scheduler'', G, Henry, Bell Laboratories
Technical Journal, Vol 63, no 8, 1984, pp. 260-272.

Also known as as ``the BSTJ 2nd special issue on Unix'', or ``BSTJ
Special Issue on Unix, vol.2'' or published with the AT&T Death-star
logo, title something like: AT&T UNIX System, Readings and
Applications, Vol. 2.  Any good CS library will have both Bell Systems
Technical Journal Special Issues on Unix, even if they don't have the
full BSTJ range.

Sure, the Fair Share Sheduler was not real-time. I dunno if the
exponential weighting between groups will work for hard-real-time
deadlines.  But it's certainly *very* well-known prior art for providing
asymmetric sharing of CPU resource across different groups of
contending processes.