Kris Pister wrote:
Touche'! It's an ugly internet out there. :)
I agree with their conclusion that "if the consequences of data
corruption are large [...] the application should add a stronger
application-level checksum."
So the truth is that I only trust networks that have 4 byte MICs at both
L2 and L4/5. Given that I know that I'm going to have that in any
networks that I build myself, 2 bytes of UDP checksum just doesn't seem
very attractive to me, and I'd like to be able to have the option to
elide them.
It is always possible disable the UDP checksum on a per-datagram basis
by setting the checksum field of the UDP header to all zeros. All IP end
hosts interpret this as a disabled checksum and do not try to compute
the UDP checksum for incoming UDP datagrams. Valid UDP checksums are
explicitly never all zeros - a zero checksum is sent as all ones instead.
A header compression scheme can safely compress away the entire checksum
field for datagrams without a UDP checksum (but leaving a bit to
indicate that the checksum is disabled).
/adam
--
Adam Dunkels <[EMAIL PROTECTED]>
http://www.sics.se/~adam/
_______________________________________________
6lowpan mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/6lowpan
