From: Chuck Lever <chuck.lever@oracle.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: quoted-printable
Message-Id: <A69DFCB1-2708-4C31-8A80-11BBA1529A2D@oracle.com>
Date: Thu, 8 Sep 2016 11:51:00 -0400
To: NFSv4 <nfsv4@ietf.org>
Mime-Version: 1.0 (Mac OS X Mail 9.3 \(3124\))
Archived-At: <https://mailarchive.ietf.org/arch/msg/nfsv4/TJuD_hW5Xo4FzueJfeV5-zpgB98>
Subject: [nfsv4] disposition of
 draft-ietf-nfsv4-rfc5666-implementation-experience
Precedence: list

Hi-

The original purpose of =
draft-ietf-nfsv4-rfc5666-implementation-experience
was two-fold:

 - To act as a problem statement for updating RFC 5666

 - To act as a stop-gap for helping future implementers of RPC-over-RDMA =
until
an updated RFC 5666 could be published

IMO these two purposes have been fulfilled, now that rfc5666bis has =
reached
WGLC.

During the March 2016 RDMA design conference call, Bill Simpson asked =
that
rfc5666-implementation-experience draft be considered to become an RFC. =
As I
understand it, he felt it contained information not in rfc5666bis that =
could
be valuable to future implementers, and thus should be made available as =
an
Informational RFC. There was no objection to publication at that time.

Strictly speaking, there's no cost to the WG for it, but publishing an =
RFC is
not free. I'm looking at my own workload and the workload of WG chair =
and the
IESG and wondering if there's another way to retain the information in =
this
document without the cost of ratification and publication.

It does not specify a protocol or even conventions that need consensus; =
all
that is already done in rfc5666bis. If it isn't, rfc5666bis should be =
changed
to include such items.

I think there is unlikely to be much change to the content of the
implementation experience document as it moves through IESG review and =
beyond.

If left to expire, this document will continue to exist in the expired =
I-D
archive. I don't see that publication lends increased value, except =
perhaps
by making its content somewhat easier to find. [*]

Lastly, is this a document that is likely to be cited by others, and =
would it
be of value to other IETF participants? I think its value is largely =
historical,
with perhaps some value as implementation advice. Very much like other =
I-Ds in
the archive of expired I-Ds.

Therefore, since it is not already cited by other I-Ds, I propose that =
this
document be dropped from the three RDMA related documents that finished =
last
call in June, and allowed to expire.

Comments?


--
Chuck Lever


[*] We have, already, another expired I-D which is part of the canon of
material recommended for RPC-over-RDMA implementers:

  https://tools.ietf.org/html/draft-hilland-rddp-verbs-00

Along with these:

  RFC 1832
  RFC 5040-42
  RFC 5531
  RFC 5666(bis)
  RFC 5667(bis)

There is some overlap between that document and RFCs 5040-42, just as =
there is
between rfc5666bis and rfc5666-implementation-experience.

Spencer Dawkins suggested creating an nfsv4 wiki where we could cite =
this kind
of reference material. Is that of interest to anyone?