On Tue, May 23, 2017 at 10:44 AM, Templin, Fred L <[email protected]> wrote: > Hi Tom, > > I am thinking about source-fragmentation only, then send all fragments with > DF=1 so they won't be further fragmented in the network. > > One use case would be for IP-in-IPv4 tunneling when there is no GUE header. > I haven't thought much about other use cases, but it should apply to any IP/X > encapsulations (X could be TCP, for example). > Introducing extension headers in IPv4 might be opening a can of worms. Could this be done in a new IPv4 option?
Tom > Thanks - Fred > >> -----Original Message----- >> From: Tom Herbert [mailto:[email protected]] >> Sent: Tuesday, May 23, 2017 10:38 AM >> To: Templin, Fred L <[email protected]> >> Cc: [email protected]; Joe Touch <[email protected]> >> Subject: Re: [Int-area] IPv6 fragmentation for IPv4 >> >> On Tue, May 23, 2017 at 9:17 AM, Templin, Fred L >> <[email protected]> wrote: >> > Joe, I wanted to run an idea by you. We all know that IPv4 fragmentation >> > has >> > problems because of the 16-bit ID field. So, why not insert an IPv6 >> > Fragment >> > Header between the IPv4 header and the upper layer protocol data, then >> > use IPv6-style fragmentation instead of IPv4 fragmentation? >> > >> Hi Fred, >> >> What is the use case for this? Unlike IPv6, IPv4 fragmentation can be >> done a routers so this technique wouldn't work in that case. If the >> fragmentation is occurring at tunnel ingress its probably just as easy >> to fragment as part of the encapsulation like we do in GUE. >> >> Tom >> >> > So, the IPv4 Protocol field would be set to '44', the IPv4 header would be >> > followed by an IPv6 Fragment Header, and the "Next Header" field in the >> > IPv6 Fragment Header would be set to the Protocol Number for the upper >> > layer protocol. Then, upon fragmentation, each fragment would have an >> > IPv4 header followed by an IPv6 Fragment Header. >> > >> > This format is shown in Figure 6 in Appendix A of the AERO draft: >> > >> > https://datatracker.ietf.org/doc/draft-templin-aerolink/ >> > >> > Does this look like something that should be broken out and put into a >> > little standalone document? >> > >> > Thanks - Fred >> > >> > _______________________________________________ >> > Int-area mailing list >> > [email protected] >> > https://www.ietf.org/mailman/listinfo/int-area > _______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
