Re: [Add] [Ext] Draft Posting: CNAME Discovery of Local DoH Resolvers

Paul Hoffman <paul.hoffman@icann.org> Fri, 26 June 2020 23:10 UTC

Return-Path: <paul.hoffman@icann.org>
X-Original-To: add@ietfa.amsl.com
Delivered-To: add@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id CC1AE3A0D03 for <add@ietfa.amsl.com>; Fri, 26 Jun 2020 16:10:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.899
X-Spam-Level:
X-Spam-Status: No, score=-1.899 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GPFSv-5Gl_vb for <add@ietfa.amsl.com>; Fri, 26 Jun 2020 16:10:49 -0700 (PDT)
Received: from ppa4.dc.icann.org (ppa4.dc.icann.org [192.0.46.77]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 121343A0D01 for <add@ietf.org>; Fri, 26 Jun 2020 16:10:48 -0700 (PDT)
Received: from PFE112-CA-2.pexch112.icann.org (out.west.pexch112.icann.org [64.78.40.10]) by ppa4.dc.icann.org (8.16.0.42/8.16.0.42) with ESMTPS id 05QNAk3o014321 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-SHA384 bits=256 verify=NOT) for <add@ietf.org>; Fri, 26 Jun 2020 23:10:47 GMT
Received: from PMBX112-W1-CA-1.pexch112.icann.org (64.78.40.21) by PMBX112-W1-CA-1.pexch112.icann.org (64.78.40.21) with Microsoft SMTP Server (TLS) id 15.0.1497.2; Fri, 26 Jun 2020 16:10:45 -0700
Received: from PMBX112-W1-CA-1.pexch112.icann.org ([64.78.40.21]) by PMBX112-W1-CA-1.PEXCH112.ICANN.ORG ([64.78.40.21]) with mapi id 15.00.1497.006; Fri, 26 Jun 2020 16:10:45 -0700
From: Paul Hoffman <paul.hoffman@icann.org>
To: ADD Mailing list <add@ietf.org>
Thread-Topic: [Ext] [Add] Draft Posting: CNAME Discovery of Local DoH Resolvers
Thread-Index: AQHWSvnIzvNV3+WnzUi6g9v0kfBTC6jr/W2A
Date: Fri, 26 Jun 2020 23:10:44 +0000
Message-ID: <0478F516-47AE-419E-B95D-F546A2F12ABB@icann.org>
References: <CABcZeBPTkWeB40wpTowKvEJ-gXA3AL2e-BE+C_FC7Js7-D0DZQ@mail.gmail.com>
In-Reply-To: <CABcZeBPTkWeB40wpTowKvEJ-gXA3AL2e-BE+C_FC7Js7-D0DZQ@mail.gmail.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: yes
X-MS-TNEF-Correlator:
x-ms-exchange-messagesentrepresentingtype: 1
x-ms-exchange-transport-fromentityheader: Hosted
x-originating-ip: [192.0.32.234]
x-source-routing-agent: Processed
Content-Type: multipart/signed; boundary="Apple-Mail=_01FD4943-D480-46B6-B782-E246539EF882"; protocol="application/pkcs7-signature"; micalg="sha-256"
MIME-Version: 1.0
X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:6.0.216, 18.0.687 definitions=2020-06-26_12:2020-06-26, 2020-06-26 signatures=0
Archived-At: <https://mailarchive.ietf.org/arch/msg/add/Zb2ukKTx5MzQqMLu-Cs9_G25spY>
Subject: Re: [Add] [Ext] Draft Posting: CNAME Discovery of Local DoH Resolvers
X-BeenThere: add@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Applications Doing DNS <add.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/add>, <mailto:add-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/add/>
List-Post: <mailto:add@ietf.org>
List-Help: <mailto:add-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/add>, <mailto:add-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 26 Jun 2020 23:10:51 -0000

On Jun 25, 2020, at 7:05 AM, Eric Rescorla <ekr@rtfm.com> wrote:

> As has been noted previously, the current Firefox DoH/TRR design
> bypasses the ISP resolver even if the ISP resolver supports DoH.
> 
> I have just posted draft-rescorla-doh-cdisco-00, which describes a
> CNAME-based mechanism for discovering when a local resolver supports
> DoH/TRR. Firefox can then determine whether that resolver is on the
> TRR list and if so can use it in preference to generic resolver.. The
> use of CNAME was chosen for pragmatic reasons (laid out in the draft).
> We're studying other designs but thought it would be a good idea
> to document this one.
> 
> See: https://www.ietf.org/internet-drafts/draft-rescorla-doh-cdisco-00.txt

A particular relevant part of this draft for this WG is:

3.2.  Why a CNAME?

   Most other proposed designs (e.g., [I-D.pp-add-resinfo] and
   [I-D.pp-add-stub-upgrade], and [I-D.pauly-add-resolver-discovery])
   use new RRtypes.  While this may be the right answer eventually, it
   is less convenient for immediate deployment, for several reasons:

   1.  It is somewhat more difficult (though not impossible) to look up
       new RRTypes on the client and provision them on the ISP resolver.

   2.  Some consumer-grade middleboxes (e.g., WiFI routers) may block
       unknown RRTypes.  The data here is quite old and limited, but
       still not particularly promising.

   The choice to use a CNAME does have one major drawback: it does not
   let us provide the URL template but only the name of the resolver.
   This is not a problem for our system in practice because Firefox will
   only connect to resolvers on a preconfigured list and thus will use
   the CNAME as a lookup key for that list.  The Mozilla team is working
   to measure the rate of new RRType interference and may revise this
   approach depending on the results of that.

   [[OPEN ISSUE: We are working to measure the rate of new RRType
   interference and may revise this approach depending on the results of
   that.]]

Point 1 feels overstated. The clients we are talking about here are modern web browsers, which can clearly write whatever they want in their DNS queries. I cannot imagine that an ISP resolver that has to also act as an authoritative server would have any restriction on what RRtypes they can provision.

Point 2 is interesting, and I'm glad to hear that Mozilla and Apple are looking into it. The key part of that is what is "unknown" to any such middlebox. The research should definitely look at whether CNAME is in the list of unknown RRtypes, given that stub resolvers rarely send CNAME queries.

The major drawback that is listed does indeed seem major for use cases other than Mozilla's TRR. This WG has been discussing other use cases, so it feels restrictive to consider a solution that is focused on just that one use case unless the middlebox problem is shown to be crippling to the use of new and more flexible RRtypes.

--Paul Hoffman