On Dec 3, 2008, at 1:10 PM, Marshall Eubanks wrote:
This would be a very encouraging reduction of inter-ISP traffic if
it could be approached in practice by what we are doing here.
This is not a simulation, but an experiment run on the Grid’5000
network with 1700 nodes. Since the lead author is subscribed to the
ALTO list, I am going to wonder publicly if it would be
possible (and sensible) to test any ALTO solution with this
experimental setup.
The only thing that is different in this approach from the standard
simulation is that a regular bittorrent client is used. The ISPs
topology, seeders to leechers ratio, upload speeds and so on are
purely artificial. The main problem with bittorrent simulations is not
the inaccuracy of the simulated software but wrong assumptions about
how the network really looks like. With respect to that, the
abovementioned experiment is not much different from "simulating
bittorrent with parameters that have unknown relation to real-world
values". Since much of the bittorrent behavior characteristics remain
unknown (although many great measurement papers have been published)
it is very hard to do credible simulations of protocol performance in
changed conditions.
AFAIK the only locality-based p2p solution, that was deployed and
tested in the real-world, with real bittorrent client, users, network
etc. is Ono. The download speed turned out to be even a bit slower
than for an unmodified client. (see figure 6a of the Ono SIGCOMM paper http://www.aqualab.cs.northwestern.edu/publications/DChoffnes08Sigcomm.pdf)
. Locality, as measured by latency, is dropping down, but it does not
boost download speed. I personally think that the actual download
speed is bounded by last-mile, not by backbone speed or connection
latency. Hence, if inter- and intradomain connections have the same
speed, traffic locality cannot give significant performance boost.
--
Maciej Wojciechowski
Delft University of Technology
_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto
