Research Of Transmission Performance Optimization Of Future Network

In information-centric network (ICN), when multiple requests simultane-ously request the same content, it will form multicast transmission because of the request aggregation mechanism. If packet loss occurs on the multicast links, as the lost packets set and the number of lost packets of each node are different, in order to complete the transmission process, the aggregation node should rec-ognize which original packets are lost and make retransmission decision ac-cording to the results. Once retransmission can only benefit few requests. Lots of pending retransmission requests will bring huge retransmission pressure to the aggregation node. This problem will become more severe especially when the number of users is tremendous and the distribution of lost packets is rela-tively decentralized, such as real time live video broadcast and video on de-mand.Motivated by the significant bandwidth efficiency improvements achieved by network coding, we propose to solve this problem with a network coding based multicast data recovery algorithm (NC-MDR) by decreasing the complex-ity and the number of data recovery of aggregation node.In this paper, we first use network coding to generate coded retransmission packets to eliminate the difference of packets. As network coding has the func-tion of information mixing, data difference will be eliminated after coded, thus when data recovery is implementing, only a certain number of coded packets are needed to send rather than some specific packets. Thus when implementing the data recovery process, the aggregation node doesn’t need to identify which packets are needed to be recovered. Cutting down the identification process de-creases the complexity of data recovery of aggregation node. To further decrease the pressure and data recovery times of aggregation node, we design a data re-covery algorithm for aggregation node.Finally, our algorithm performance analysis and simulation verification of NC-MDR are performed on ndnSIM. First, we give out the simulation results of chunk size problem which is-mentioned in Chapter 3 and obtain the most suita-ble value of chunk size from simulation results. Then simulation results of computation time overhead of network coding are also displayed. At last, we se-lect the number of retransmissions of aggregation node and the flow completion time to be the critical parameters of performance simulation, and comprehensive simulation verification of NC-MDR algorithm are carried out under two differ-ent scenarios to demonstrate the effectiveness of NC-MDR algorithm.