Hi Joe, > -----Original Message----- > From: Joe Touch [mailto:[email protected]] > Sent: Tuesday, February 24, 2015 10:12 AM > To: Templin, Fred L; Ronald Bonica; [email protected] > Subject: Re: [Int-area] draft-ietf-intarea-gre-ipv6 > > > > On 2/24/2015 9:35 AM, Templin, Fred L wrote: > > Hi Joe, > ... > >> We've been around the block on "let's describe what SHOULD be, but isn't > >> deployed". While I agree that's important, that is not the function of > >> this document. > > > > There are two different intertwined issues that need to be teased apart. > > > > First issue is that this document and RFC2473 are in the same boat in terms > > of Fragmentation and MTU and the former should therefore cite the latter. > > I do agree on this point. > > > Or, if the former has more to say (e.g., sending PTB with a size less than > > 1280) then it should also update the latter. > > That doesn't follow. This doc is about IPv6; it can surely augment the > basic rules in another PS that describes an IPv4 mechanism without > "UPDATING" that doc (i.e., it need not apply to IPv4).
RFC2473 is about generic packet tunneling over IPv6 (not IPv4). > > Second issue is that both documents are susceptible to black holes if > > PTB messages are lost. > > That may be good to point out, but it does not argue for changing what's > in this doc. We have many protocols that are either incomplete or have > land mines built-in; they are widely deployed and there's nothing wrong > with documenting *what is deployed*. This document is going for proposed standard so, as Ron inferred, they are trying to document "what is and what will forever be". > When we do document what is deployed, we need to avoid addressing how to > fix those issues, though. If the document is suggesting a long-term solution (which it seems to be, according to Ron), then it needs to address how to fix the issues. Thanks - Fred [email protected] > Joe _______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
