>>>>> On Tue, 05 Mar 2002 12:56:58 +0100,
>>>>> Kristof Verhenne <[EMAIL PROTECTED]> said:
> I understand that when packets are too big to reach their destination,
> because mtu is smaller than a packet they are discarded and an
> icmpv6-packet too big- packet is sent to the source.
> Now, when would one ever want to use a fragmentation header?
> I read that fragmentation only can be used by source nodes. Why
> wouldn't they want to use the flowlabel instead of an extra header?
Because it would be difficult to encode all the parameters in a
fragment header (offset, ID, the M bit) into a 20-bit flowlabel.
Also, without a fragmentation header, we cannot separate the
"unfragmentable part" from the "fragmentable part" in each fragment.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
---------------------------------------------------------------------
The IPv6 Users Mailing List
Unsubscribe by sending "unsubscribe users" to [EMAIL PROTECTED]
