Subject: Re: Throttling IO Requests in NetBSD via Congestion Control
To: Erik Berls <>
From: Sumantra Kundu <>
List: tech-kern
Date: 12/09/2006 19:13:00
> What is the granularity of the queues?  Is it for the whole system? by disk?
> by filesystem?
The queue (read/write) resides inside the IO scheduler. Requests are
clubbed together per-disk (think of it as virtual queue) so as to
identify the dynamics of read/write per disk. The algorithm works on a
per-disk basis.