online read us now
Paper details
Number 1 - March 2016
Volume 26 - 2016
A mathematical model for file fragment diffusion and a neural predictor to manage priority queues over BitTorrent
Christian Napoli, Giuseppe Pappalardo, Emiliano Tramontana
Abstract
BitTorrent splits the files that are shared on a P2P network into fragments and then spreads these by giving the highest
priority to the rarest fragment. We propose a mathematical model that takes into account several factors such as the peer
distance, communication delays, and file fragment availability in a future period also by using a neural network module
designed to model the behaviour of the peers. The ensemble comprising the proposed mathematical model and a neural
network provides a solution for choosing the file fragments that have to be spread first, in order to ensure their continuous
availability, taking into account that some peers will disconnect.
Keywords
P2P model, neural network, wavelet, diffusion, file sharing