[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [RRG] Are we solving the wrong problem?
Hi Scott,
El 05/03/2008, a las 17:23, Scott Brim escribió:
Excerpts from Pekka Nikander on Wed, Mar 05, 2008 07:39:36AM +0200:
Interesting. This sounds pretty much like HIP base exchange...
I had exactly the same thought, smiling as I read.
This is interesting stuff but mostly orthogonal to routing (alas).
I am not yet sure about that (i am not sure that is not the case
neither though)
It would be possible to think that TCP marks the packets saying that
it wants a different path, cause the currnet one is congested, and
then we would need a much richer interaction with routing.
From my perspective, this is not routing but it is not transport
neither, it is the interaction between them and there is interesting
and promising work to be done there imho
It
allows endpoints to manage their congestion and do simple TE.
I disagree with the word simple :-)
I think this provides really interesting TE because is quite dynamic
TE response to congestion. this is what people try to achieve today by
injecting more specifics that are highly volatile. More specifics for
TE represrent about 30% of current routing table (Huston, Zhang et
al). Moreover, i would think thjat the represent the most volatile
ones, since they are likely to change because of the conditions, so
the are likely to account for an imporntat part of the updates. Maybe
doing this type of mechanism would allow automatic rerouting to avoid
congestion and will avoid the more specifics related to that (of
course, the more specifics related to a different type of TE needs
wouldn't go away)
It
doesn't allow more sophisticated policy, avoid renumbering, etc. (see
the RRG design goals).
agree that some other forms of policy routing are not so clear, but
note that the ISPs can give congestions hints to make the flows to be
rerouted, so it would be indeed possible to achive some forms of
policy manipulation with this
renumbering would need more work, and doesn't seem to me to be a key
feature of this approach to put in some way :-)
regards, marcelo
--
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
--
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