[bittorrent] New version of the paper "Rarest First and Choke Algorithms Are Enough "
Arnaud.Legout at sophia.inria.fr
Fri Jun 2 12:00:55 EDT 2006
I received a lot of good comments on the first version of this paper.
You can find a new version of the paper
that takes into account these comments here:
The major modifications compared to the previous version are:
-New torrents with a single (or few) seed(s) and several
hundred of leechers were added. These new torrents are now used as an
illustration of the transient and steady state. This is to answer to the
comments that pointed out that small torrents are not challenging
for the rarest first algorithm.
-I rewrote a large part of the methodology section in order
to better discuss the experimental setup and the limitations of the
study. In particular I discuss the single client monitoring restriction,
and how the torrents were chosen.
-A new figure is presented for the section on the choke
algorithm to better illustrate the correlation between
the number of unchokes and the interested duration.
As always, comments and critics are welcomed.
More information about the BitTorrent