Hi,
On 1/27/16 4:50 PM, Ted Hardie wrote:
Recent proposals discussed in the IETF have identified benefits to
more distinctly identifying the hosts that are hidden behind a shared
address/prefix sharing device or application-layer proxy. Analysis
indicates that the use of a TCP option for this purpose can be
successfully applied to some use cases.
The proposals have identified benefits according to the authors,
but the IETF declined to adopt them because there were far bigger
downsides. This abstract gives the opposite impression, which is not
exactly kosher.
I agree with Ted about the abstract; that at the very least it should be
reworded. Referring to the IETF needs to be fairly done in the
independent series. We already have enough problems with people
distinguishing IETF from independent work. However...
Armchair lawyers will also note that the procedure refers to "IETF
specifications". An independent-submission RFC is *not* an IETF
specification.
Yes, and we have historically said that publishing things in
the ISE stream when they counter IETF specifications can
only be done when the IESG deems there to be no conflict
with IETF specifications. This clearly does conflict with the
thrust of efforts in the IETF (e.g. tcpinc).
RFC 5742 is quite explicit. BCP88 itself is not a specification but a
Best Current Practice, and limits its applicability to IETF, and not
independent, specifications. As to whether this conflicts with "the
thrust of efforts in" tcpinc, that is not what the IESG wrote. If it
had written it, the justification would have to again be based on what
is stated in RFC 5742, specifically in Section 3. There should be at
least sufficient information to indicate what the nature of the conflict
is. Having a different opinion than that of the IETF or a working group
is *not* a justification for non-publication, but rather a justification
for the existence of the ISE.
In short, I believe the IESG erred procedurally, and would suggest they
revisit their approach. Haralds query about attaching an IESG note
seems entirely appropriate.
Eliot
signature.asc
Description: OpenPGP digital signature