In a content distribution (file sharing) scenario, the initial phase is delicate due to the lack of global knowledge and the dynamics of the overlay. An unwise piece dissemination in this phase can cause delays in reaching steady state, thus increasing file download times. After showing that finding the scheduling strategy for optimal dissemination is computationally hard, even when the offline knowledge of the overlay is given, we devise a new class of scheduling algorithms at the seed (source peer with full content), based on a proportional fair approach, and we implement them on a real file sharing client. In addition to simulation results, we validated on our own file sharing client (BUTorrent) that our solution improves up to 25% the average downloading time of a standard file sharing protocol. Moreover, we give theoretical upper bounds on the improvements that our scheduling strategies may achieve.
and is available at : http://dx.doi.org/10.1016/j.comnet.2011.06.020