Problem
Consider transmission of voice over the internet using the following settings/protocols:
1) VoIP codec used: G.711 (64,000 bps). Sent inside RTP packets, with the shortest possible 12-byte headers.
2) Voice frames are transmitted every 10 ms using a UDP transport-layer protocol, with 8-byte headers
3) UDP datagrams are then sent using IPv4 (20-byte headers)
4) Finally, IPv4 packets are transmitted using Ethernet frames (38 bytes used by header and trailer structures)
What is the maximum overall efficiency of such a setup? I.e., what percentage of the transmitted Ethernet frames is going to be comprised of the actual voice data, as opposed to the various overhead information? Show your calculation.