[lkml]   [1996]   [Jun]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: 2.0, loggings, cpu quotas, 2.1 issues, etc.
One (too) simple choice might be to charge for kernel time .. CPU spent
swapping for a process is added to CPU load from that process. If CPU % is
then quotad it will keep their process from causing more than it's percent
of system load.

While this falls out of the fairness test in many ways, it does provide the
incentives all lined up in the right direction: towards non-peak times, low
memory consumption and sharing of resources. (run a script that watches the
atime on the library you want :) )

I assume any CPU dividing is ignored when only one process is ready-to-run?

Also this would cut up a lot of denial-of-service attacks.

For example a program which allocated (physical memory)+1 bytes and then
poked around at random would mostly cripple itself by queueing up swapping
time on the counter -- but then you would have to divide up the "idle" time
the kernel spends while all the processes are blocked waiting for the swap

Ok so I didn't say it would be fair. (shrug)

You have new mail in /dev/null

 \ /
  Last update: 2005-03-22 13:37    [W:0.119 / U:4.476 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site