[saag] Re: QUIC-ify everything?

Benjamin Kaduk <kaduk@mit.edu> Sun, 18 October 2026 05:40 UTC

Received: from outgoing.mit.edu (outgoing-auth-1.mit.edu [18.9.28.11]) (using TLSv1.2 with cipher ECDHE-ECDSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mx.ietf.org (Postfix) with ESMTPS id BB83746 for <saag@ietf.org>; Sun, 18 Oct 2026 05:40:27 +0000 (UTC)
Authentication-Results: mx.ietf.org; dkim=pass header.d=mit.edu header.s=outgoing header.b=IFMKEblx; spf=pass (mx.ietf.org: domain of kaduk@mit.edu designates 18.9.28.11 as permitted sender) smtp.mailfrom=kaduk@mit.edu; dmarc=pass (policy=none) header.from=mit.edu
Received: from kduck.mit.edu ([73.151.147.203]) (authenticated bits=56) (User authenticated as kaduk@ATHENA.MIT.EDU) by outgoing.mit.edu (8.14.7/8.12.4) with ESMTP id 69I5eESs013588 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Sun, 18 Oct 2026 01:40:17 -0400
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=mit.edu; s=outgoing; t=1792302018; bh=k6bDO6bIjzZDA51ooyqRTEdIUgQHXn2v0PcE8uj3x+c=; h=Date:From:Subject:Message-ID:MIME-Version:Content-Type; b=IFMKEblxxTUv7hdW6wGWOoXqriKoE9LC32pCUEOXqXjoR19Hs4JRSx9dTHdRldFnX CFPyToPP6aqtEeSJ68Y5qfeZend9EzrHCi2BG3lJS48Jwh+agXT0a0b7+SWOp94P+d bpsG2gmSPm3zZ9uzNsQ2oU5aMOqM6HIbRyzcFGqqHa5X2lQTZOKs0B355ai1HiCdzC 9adhb+G2z0d44ElF3fgZXr5++gAXNPoaZhA0r8KyzIwbt8xX955JVTTPawFSawCzZN KjbQPI9u8th83wgx1jzVV/xsBqBB7bXnYnyhDWVLalOhByQgujatsVGZwAoB21Z9+5 KoRb1vvMPXWmA==
Date: Sat, 17 Oct 2026 22:40:13 -0700
From: Benjamin Kaduk <kaduk@mit.edu>
To: Michael Richardson <mcr+ietf@sandelman.ca>
Message-ID: <atRbvVNiHsN0zeAo@kduck.mit.edu>
References: <CAMm+LwhkeRmTcKKijWfYDRT+m52F_zqkrJaZakb4yZ8QD6-CiA@mail.gmail.com> <1077284.1792254378@dyas>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Disposition: inline
In-Reply-To: <1077284.1792254378@dyas>
X-Spamd-Bar: -------------------
Message-ID-Hash: UAWCOBA3SZVJWSWEE62UCS5F7ZA57DIZ
X-Message-ID-Hash: UAWCOBA3SZVJWSWEE62UCS5F7ZA57DIZ
X-MailFrom: kaduk@mit.edu
X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; loop; banned-address; header-match-saag.ietf.org-0; header-match-saag.ietf.org-1; emergency; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header
CC: IETF SAAG <saag@ietf.org>
X-Mailman-Version: 3.3.10
Precedence: list
Subject: [saag] Re: QUIC-ify everything?
List-Id: Security Area Advisory Group <saag.ietf.org>
Archived-At: <https://mailarchive.ietf.org/arch/msg/saag/NVUuXOCDwM9ZTI5rd3HCQrUBZa4>
List-Archive: <https://mailarchive.ietf.org/arch/browse/saag>
List-Help: <mailto:saag-request@ietf.org?subject=help>
List-Owner: <mailto:saag-owner@ietf.org>
List-Post: <mailto:saag@ietf.org>
List-Subscribe: <mailto:saag-join@ietf.org>
List-Unsubscribe: <mailto:saag-leave@ietf.org>

On Sat, Oct 17, 2026 at 06:26:18PM +0200, Michael Richardson wrote:
> 
> Phillip Hallam-Baker <phill@hallambaker.com> wrote:
>     > I have just mounted a share from my internal NAS on the proxmox
>     > host and can view the files on it from that host and at no point
>     > did I give an account, password or any other form of
>     > authentication. Is the state of the art really so lame? Am I seeing
> 
>     > All it would take to create secure NFS would be to run the RPC
>     > calls over QUIC.
> 
> That's not really the hard part.
> The hard part is mapping the credential used to a file system (Unix ID) on
> the server.   And you might want more than one mapping: N:M

Yes, identity-management is the hard part.  Because your NFS-protocol
client (i.e., host) probably only has one identity, but may (in the general
case) have more than one user, and then your identity mapping problem gets
a lot more exciting.

Over in AFS-land, we embraced this separation of identities and let you
provide both host credentials and user credentials to get a combined
identity that is what you present to the fileserver [1], which can then
apply whatever authorization policy it wishes (just defaulting to per-user
credentials may be simplest, and the system mostly works fine even in the
absence of host identity credentials).  I'm told that Auristor, Inc [2] has
done some interesting things in this space, but I hvae not explored it
myself out of IP considerations.

-Ben

[1] draft-wilkinson-afs-rxgk-afs is where this ended up
[2] https://www.auristor.com/