Pekka Savola wrote:
On Wed, 11 Oct 2006, Jari Arkko wrote:
- 'Fragmentation Considered Very Harmful '
<draft-heffner-frag-harmful-02.txt> as an Informational RFC
I raised this in private comment, but I'll mention this again because
the authors disagreed with the first point and apparently forgot the
second one. Hopefully this would result in others taking a look at the
draft as well.
- The title should really be "IPv4 Fragmentation Considered Very
Harmful" plus some minor wording changes. The document does not as
written provide sufficient justification why IPv6 fragmentation should
be considered "very harmful" as IPv6 has 2^16 more bits of fragment
space, mandatory PMTUD, and routers not fragmenting packets.
Someone has also just pointed out a potential ambiguity in the term
"Fragmentation" so we may be changing the title.
[snip]
- in a couple of places like in 2nd to last paragraph of S2, and S6, you
forget to mention that the protocol number is also a disambiguating
factor in the fragment ID generation.
Yes, this was an oversight and has been corrected in our working copy.
(I think S6 was fixed in -02, please correct me if I'm wrong.)
Thanks,
-John
_______________________________________________
Int-area mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/int-area
