[p2p-hackers] network coding put to practical p2p content dist. use

coderman coderman at gmail.com
Mon Jun 20 18:16:34 UTC 2005


On 6/19/05, Nick Johnson <arachnid at notdot.net> wrote:
> How is the use or otherwise of a tracker in any way related to the use of
> rateless codes?

coordinating exchange of authentic blocks of coded content is the crux
of the issue.  if you have centralized control, error / online coding
does not get you any benefit over the current method of using verbatim
blocks of data.

in a decentralized context, meaning no centralized tracker, the issue
of authentic coded blocks is difficult and conveniently not described
in the network coding paper.  if you can do this efficiently then
there is potential that network coding (with each peer performing
re-encodes with current data) can be efficient without requiring the
centralized control that is critical to the current unencoded block
transfer coordination in bittorrent.



More information about the P2p-hackers mailing list