RE: queueing and priorizing strategies

On Oct 15 2007, Philippe Verdy wrote:
>My opinion is that this will not work reliably: others can lie.

The original idea was that you'd directly measure the amount downloaded 
from and uploaded to each peer (as in BitTorrent), so there'd be no way for 
a peer to lie. Arne's suggestion (sharing information about the amount 
uploaded and downloaded) can be made secure against liars by using a 
Sybilproof reputation mechanism based on maxflow - see the paper linked in 
my previous mail for details.

Cheers,
Michael