A path selection approach with genetic algorithm for P2P video streaming systems

Abstract

In peer-to-peer (P2P) video streaming applications, the Quality of Experience (QoE) obtained by peers may vary according to the level of network congestion on the path between the peer and its parents. In this paper, we have proposed a parent selection approach for hybrid push-pull based video streaming systems, taking physical paths into account. In this approach, we proactively measured the disjunction of paths between peers and partners and chose the set of parents among partners having maximum disjoint paths by using a genetic algorithm at the beginning of the streaming session. The peers continue the paths of parent set to keep disjoint during streaming. The proposed system was tested on NS3 and the obtained results show that the proposed approach provides an increase up to 10 % in continuity index compared to hybrid push-pull based system which does not consider underlying network infrastructure. Furthermore, higher performance in terms of other parameters related to QoE such as Peak Signal to Noise Ratio (PSNR), reset count, and network related parameters such as video propagation success among peers and re-selection rates of new parents is observed with the proposed approach.