Optimized Packet Scheduling for Live Streaming on Peer-to-Peer Network with Network Coding

This paper proposed an optimized packet scheduling algorithm for live peer-to-peer streaming system, where network coding technique is extended to improve the efficiency in bandwidth utilization. We identify a problem of undesirable non-innovative packet transmission due to the latency of buffer-map update among peers, which, with many previous proposed techniques, yields to bandwidth inefficiencies. In the proposed system, we present an optimized packet scheduling algorithm for forecasting the number of required packets at parent nodes in advance, a selection mechanism for selecting and encoding packets for forwarding, and an adaptive push algorithm for smart frame skipping. The proposed optimized packet scheduling algorithm calculates the minimum number of required packets for full rate transmission. Transmitting as the scheduled results can reduce non-innovative packet transmission, save limited bandwidth for innovative transmission, thereby improving the streaming continuity. The simulation results show that the proposed scheme provides significantly better video quality, delivery ratio, lower redundancy rate, and higher innovative video packet rate compared with previous packet scheduling algorithms.