On 04/30/2011 11:10 AM, Richard Alimi wrote:
Very interesting results! It does look like it converges to a good
result, but I'm wondering if it would be possible to reduce the
convergence time further (smaller than 10m). Can you give some more
details on how ALTO was integrated with the BitTorrent clients here?
Rich: Sure. Please see inline.
Was it used for ranking peers returned from a tracker, or also
peer-exchange (if enabled)?
We have not implemented peer-exchange in our ns-3 model. So,
the results you see are only from getting the peer list from a
tracker.
How often were new peers discovered from a tracker or peer-exchange?
A peer in our model contacts the tracker every 30s of simulated time.
Was it a greedy strategy in which only closest peers were chosen, or
were a few sub-optimal peers chosen for robustness?
For starters, the peer list returned by the tracker is not modified
to remove any peers; the peer list is simply re-ordered with the
least desirable peers appearing at the bottom of the list.
The announcing peer gets this list from the ALTO server and starts
to contact each peer in the list starting from the top of the
list, subject to the constraints that:
(a) it may already have an existing connection with that peer,
so no point in establishing a new one;
(b) the contacted peer may not accept the announcing peer's connection
request;
(c) the contacted peer may accept the announcing peer's request, but
choke it until further notice;
Thus, a few sub-optimal peers will sneak depending on the constraints
outlined above. I have not done an analysis on output of the
simulation to see what is the distribution of the connections between
an announcing peer to sub-optimal peers, but intuitively, it appears
that there will be a minority of suboptimal peers.
Thanks,
- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / [email protected]
Web: http://ect.bell-labs.com/who/vkg/
_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto
