Optimizing random network coding for multimedia content distribution over smartphones

Random Linear Network Coding for 5G Mobile Video Delivery

“…However, in this case, RLNC provides similar performance as compared to already proposed AL-FEC solutions such as Raptor codes [57] or LDPC triangle/staircase codes [58], providing limited benefit since no re-encoding at intermediate nodes is typically assumed in multicast/broadcast setup. Nevertheless, a large number of academic studies considered application and optimization of RLNC for mobile multicast video delivery demonstrating various benefits in different scenarios [59][60][61][62]. Besides notable benefits of AL-RLNC in terms of flexibility and ease of implementation, there are several drawbacks of AL-RLNC worth emphasizing.…”

Energy Efficient QoS-Aware Random Network Coding on Smartphones

“…A number of works have considered ways to improve the throughput and reduce the decoding latency of RNC [3][4][5][6][7][8][9][10][11][12][13]. RNC was proposed by Ho et al [1] as a capacity-achieving scheme for multicasting in wired networks following the work by Ahlswede et al [2] who first suggested network coding and showed its utility in wireline networks.…”

“…Several previous works have investigated the performance of RNC on smartphones [13,[17][18][19][20]. Shin et al in [13] showed that parallel processing techniques such as multithreading and SIMD can achieve at least an order of magnitude throughput improvement compared to previous results as reported in [17][18][19][20].…”

“…Shin et al in [13] showed that parallel processing techniques such as multithreading and SIMD can achieve at least an order of magnitude throughput improvement compared to previous results as reported in [17][18][19][20]. SIMD instructions were used to speed-up a loop-based GF multiplication method in [3] whereas in [13] a table lookup method was used in combination with SIMD. Approaching differently to RNC's high decoding latency problem, a variant of RNC called Pipeline Network Coding (PNC) has been proposed in [21].…”

“…To overcome the computational overhead of RNC, a number study focused on the performance of RNC [3][4][5][6][7][8][9][10][11][12][13]; recent advances have succeeded in mitigating the concerns. According to one recent study [13], the throughput of an optimized RNC decoder on smartphone platforms is over 1 Gbps, which well exceeds H.264 1080p30 FHD bandwidth requirement of 30 Mbps. This poses an interesting question: is there any resource waste in outpaced RNC decoding?…”