IP V6 overhead

[Francis Dupont]

 In your previous mail you wrote:

   Secondly I wonder if someone can explain to me the following:
   We are using the IP network for IP Telephony (VoIP) . Currently we are
   doing it by using IPV4
   At IPV4 we are sending packet every 20 ms. The length of the payload is 160
   bytes (without compression) and the overhead is 78 bytes (IP, UDP, RTP, L2
   & L1). With IPV6 the overhead will be much larger because of the IP address
   fields.
   So we will reach a situation where the overhead will be much larger than
   the payload and the required bandwidth will be increased.
   How such a problem is solved at IPV6 ?

=> with a good header compression you can remove most of the headers (*).
Look at RFCs 250[789], RFC 3095 and future ROHC RFCs...

Regards

Francis.Dupont@enst-bretagne.fr

PS: in the IPv6 header everything can be compressed (this is not true
for IPv4) so the only needed thing is the index of the context which
takes the base 2 logarithm of the number of contexts. On a perfect
point-to-point link with only one VoIP conversation the IPv6 overhead
is *zero* bits (sorry, can't do better :-)!