[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Diffserv] Model - queues

> > What you're pointing out is that the description doesn't capture the notion of
> > a trigger, which could either be fired by a packet arrival at the discarding
> > element, some state change in the FIFO element, scheduler, or meter, or a
> > timer.  That's useful.
> And also that the discarding decision can apply to any packet in the
> buffer,
> not necessarily the arriving one or the one at the tail/end of a linked
> list.
> You could pick a random one (not sure why, though ;-) or actually more
> than
> one, e.g., kicking out several small packets to make room for a big one.
Two separate issues: if you wanted to allow picking a random packet to discard, then the data structure is no longer a FIFO.  We could do that, but since you're not sure why, and I'm not sure why, I'd just as soon not unless somebody else comes up with a reason why.  The proposal doesn't specify how many packets a discarding element can drop on a single trigger;  my assumption was 'zero or more'.  [As an aside, I'd be nervous about kicking out several small packets to make room for a large packet, since if the small packets are  TCP ACKs, you'd start getting into ack compression and slow start.] 

diffserv mailing list
Archive: http://www-nrg.ee.lbl.gov/diff-serv-arch/