Some quick notes: RFC1191 is for IPv4. The RFC for IPv6 PMTUD has just been revised (8201), and includes clarified and corrected terms for the various types of MTU that need to be considered.
It might also be useful to look at draft-ietf-intarea-tunnels and its terminology for MTU values, esp. at the link layer and considering link fragmentation and reassembly. RFC5405 similarly is out of date. The authors may want to consult https://www.isi.edu/touch/tools/rfc-what-i-mean.html (it won't catch citing RFC8201 instead of RFC1191 because the former is IPv6 vs the latter for IPv4, rather than "obsoleting" the other). The issue of overlapping fragments should be addressed (permitted or prohibited). The issue of needing to have the IPv6 header should be more clear about this also necessarily including any HBH EHs, because it would be inappropriate to route solely on the base IPv6 header if there are applicable HBH EHs that need to be considered. I'm not sure about the conclusion of the security considerations - esp. given the MAC mechanism in the preceding section, which appears to create an attack vector (AFAICT). FWIW. Joe On 7/12/2017 12:23 PM, Spencer Dawkins at IETF wrote: > Dear Transporters, > > Gab Montenegro, co-chair for 6lo, let me know that 6lo is considering > adoption of a draft about how 6lowpan forwards fragments. > > The draft > is https://datatracker.ietf.org/doc/draft-thubert-6lo-forwarding-fragments > <https://datatracker.ietf.org/doc/draft-thubert-6lo-forwarding-fragments>. > > Gab asked about getting TSV eyes on this draft, to help inform the > working groups conversation about adoption. > > 6lo is scheduled for 15:50-17:50 on Tuesday Afternoon (session II), > opposite MPTCP, but if you're available, informed, and opinionated, > Gab says the working group has time to discuss adoption at the end of > their agenda, and welcomes input. > > Thanks! > > Spencer, who also plans to attend, but is not the guiding light in TSV > for fragmentation issues ...
