[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [RRG] ALT's strong aggregation often leads to *very* long paths
> From: Brian E Carpenter <brian.e.carpenter@gmail.com>
> Regardless of the details, is it a surprise that the path taken by
> initial packets while waiting for a mapping have significantly greater
> length than the path established after the mapping arrives?
Except in full-push systems like NERD...
> There seems to be an inevitable design choice between:
> 1. Discard those packets;
> 2. Queue them;
> 3. Send them a long way round.
Excellent point. There are obviously engineering implications behind doors #2
(queue management) and #3 (hot-spots at the top of the resolution hierarchy,
if you do some sort of hierarchy-based alternative forwarding), but that's to
be expected.
> All these choices are compatible with the datagram model, so I don't
> see how any of them can be architecturally "wrong".
"The difference between theory and practise is even bigger in practise than it
is in theory." :-) Yes, in theory one could chose door #1, and just toss them,
but in the real world people don't seem to like that one.
> From an engineering viewpoint, #3 seems reasonable (i.e. this minor
> delay will be added to the other delays for starting an
> application-level session).
Yes... which is why a number of ideas have been floated for how to do that
(multicast groups, forwarding along the resolution tree, etc). I wonder if
there are any good ideas we haven't explored yet on how to do this?
Noel
--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg