From: John Gibson (johngibson1541@yahoo.com)
Date: Wed Apr 25 2007 - 11:27:50 ART
We know routing engines look at the "do not fragment"
bit of the packet. That is in IP header, irrelevant to
switching engines.
I am reading 6500 series document "Configuring Jumbo
Frame Support" section. It does not say any thing
about fragmenting frames.
IPv6's rfc says all routers claiming IPv6 support are
not allowed to fragment any IPv6
packets equal or smaller than 1500 bytes. And
this implies if an interface's MTU is smaller than
1500, it needs to fragment the frame , not the packet.
Fragmented frames are supposed to be reassembled at
the very next routing engine? Who knows. Fragmented
packets are reassembled at the receiver end host. We
know this for sure. Serial interfaces can have MTU
smaller than 1500 right? But serial lines are point
to point , no such thing as switching or bridging.
Can a frame relay switch be configured with a MTU
different from the DTE router's serial interface's
MTU. I will assume not.
So, when IPv6's rfc talks about fragmenting frames,
the scenario doesn't exist in our current world.
I am going to get the CCIE without thinking about
some bridging or switching equipment
in this world that support MTU smaller than 1500.
This archive was generated by hypermail 2.1.4 : Tue May 01 2007 - 08:28:37 ART