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

Re: [p2prg] New version of the mythbuster-ing draft



Enrico,

The savings will indeed have something to do with "the number of peers per
AS" but this can be refined.

To save cross-domain traffic "globally," you need 1) domains to be big
enough so you will have more than one *simultaneous* peer for big torrents
and that 2) the number of big torrents per domain to outweight the number
of small torrents. Even if it's possible to save a lot for big torrents,
that doesn't mean you can save for most torrents, especially because there
are more small torrents than big ones. Fortunately though, big torrents
also generate much more cross-domain traffic than small ones.

As I said during my talk at the last IRTF session, we have found that the
top 1,000 torrents on the 214,443 torrents that we have crawled generated
about half of the "global" cross-domain traffic.

Let me know if this answers your question or if you want me to elaborate
more about this topic. You can also refer the Section 6.3 of our paper
[LeBlond] for more info.

> How about replacing the above with the following text?
>
> 4.  Experiments with real BitTorrent clients run by researchers a INRIA
>     [LeBlond] have shown a 40% reduction in cross-domain traffic on a
>     global scale, with local peaks up to the 99.5% in exceptionally
>     favorable conditions.
>

That's better but I would add some numbers to make it clear that "global
scale" actually means at the scale of the Internet. "Global" scale could
be interpreted as all torrents (resp. a single torrent) but from the point
of view of a single AS. Something like this:

4. Experiments with real BitTorrent clients and real distributions of peers
   per AS run by researchers at INRIA [LeBlond] have shown that ASes
   with 100 peers or more can save 99.5% of cross-domain traffic
   with high values of locality.

   They have also shown that at a global scale, i.e., 214,443 torrents,
   6,1113,224 unique peers, and 9,605 ASes, high locality can save 40% of
   global inter-AS traffic , i.e., 4.56 Petabytes (PB) on 11.6 PB. This
   result shows that locality would be beneficial at the scale of the
   Internet.

Stevens