Not sure why - you asked, AFIACT, how BCPs for link design affect things that don’t design links (NATs).
I confirmed that they don’t. Was there a different question? Joe — Joe Touch, temporal epistemologist www.strayalpha.com > On Dec 8, 2021, at 12:39 PM, Templin (US), Fred L <[email protected]> > wrote: > > Joe, I am having a hard time seeing your response as anything other than a > non-answer to my question. > > Fred > >> -----Original Message----- >> From: [email protected] [mailto:[email protected]] >> Sent: Wednesday, December 08, 2021 12:11 PM >> To: Templin (US), Fred L <[email protected]> >> Cc: Dino Farinacci <[email protected]>; [email protected] >> Subject: Re: [Int-area] Side meeting follow-up: What exact features do we >> want from the Internet? >> >> Hi, Fred, >> >>> On Dec 8, 2021, at 11:52 AM, Templin (US), Fred L >>> <[email protected]> wrote: >>> >>> Joe, RFC3819, Section 2 in particular gives BCPs for setting link MTUs. >> >> NATs and tunnels don’t have control over the link MTUs over which they >> operate; the user doesn’t have control over how those are >> composed or interact. >> >>> By my read, the >>> only links that would set an MTU smaller than 576 should therefore only >>> occur at the >>> network edges; not somewhere in the middle of the network. >> >> Tunnels create an tunnel MTU (which is the link MTU, thinking of the tunnel >> as a link) by fragmenting at the ingress and reassembling at the >> egress. >> >> That happens anywhere in the network. While it PRESENTS an effective MTU of >> the tunnel as a link, it doesn’t operate as if it avoids >> fragmentation. >> >> Joe >
_______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
