Network Coding in Ad Hoc Networks: A Realistic Simulation Study

Adaptive Random Linear Network Coding with Controlled Forwarding for wireless broadcast

“…As the number of nodes increases, there are more chances to successfully decode the packets, but on the other hand a conflicting situation arises as there are more collisions and the chances of a decoding failure increases as well. Their P loss results are also in conflict with the results mentioned in [15], who showed that RLNC shows poor performance when the number of nodes increases. Another result between P loss and length of coding vector shows that there is strong influence of the size of the coding vector and P loss .…”

“…As an example [9] used 40 kbps and [15] used 50 kbps for single source scenarios. It is mentioned in [15] that their RLNC based probabilistic broadcast algorithm has high packet loss for higher data rates. We used up to 100 kbps for a single source and up to 25 kbps per source for 04-source scenarios.…”

“…If all the packets of a particular block are received, then they are decoded and recovered. However, as mentioned in [15], some of the packets can be recovered even if fewer than blocklength packets are received if the rank of the sub matrix is full. However, this paper did not talk about decoding based on partially received packets.…”

“…In [15], the authors observed the effect of packet loss and propagation delay using RLNC. Their work addresses single-source broadcast in wireless adhoc networks.…”

“…Secondly, decoding failure causes all packets to be lost. They did not look at the possibility of earlier decoding of packets as mentioned in [15]. The metric used to judge the performance is decoding failure, packet loss probability vs. node density, vs. length of coding vector N and vs. order of GF.…”