[Snac] Re: 2nd WGLC: draft-ietf-snac-simple (ends 2026-09-23)

Esko Dijk <esko.dijk@iotconsultancy.nl> Wed, 21 October 2026 11:03 UTC

Received: from dane.soverin.net (dane.soverin.net [185.233.34.21]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange x25519 server-signature ECDSA (prime256v1) server-digest SHA256) (No client certificate requested) by mx.ietf.org (Postfix) with ESMTPS id 9E0A746 for <snac@ietf.org>; Wed, 21 Oct 2026 11:03:35 +0000 (UTC)
Authentication-Results: mx.ietf.org; dkim=pass header.d=iotconsultancy.nl header.s=soverin1 header.b=BsoUBvAW; spf=pass (mx.ietf.org: domain of esko.dijk@iotconsultancy.nl designates 185.233.34.21 as permitted sender) smtp.mailfrom=esko.dijk@iotconsultancy.nl; dmarc=pass (policy=reject) header.from=iotconsultancy.nl
Received: from smtp.soverin.net (unknown [10.102.10.155]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange x25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) by dane.soverin.net (Postfix) with ESMTPS id 4j8mc63C2Kz9t; Wed, 21 Oct 2026 11:03:34 +0000 (UTC)
Received: from smtp.soverin.net (smtp.soverin.net [10.102.10.155]) by soverin.net (Postfix) with ESMTPSA id 4j8mc60g7yzfjG; Wed, 21 Oct 2026 11:03:34 +0000 (UTC)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=iotconsultancy.nl; s=soverin1; t=1792580614; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: in-reply-to:in-reply-to:references:references; bh=Gm2wawgdHh4VWclhSKCNUFUSn54M41icp5MFHw2MjDY=; b=BsoUBvAWBYo2hWBTCvx5LASqobAI4pnd3zvBtcvoFfO9ggG3Z/R/8HCDAkY35VR4Nx7WeD ZromE6t7B3Lg9TmzcNSPKshyUPRhetRGi8TSN4i4ecvhqiQ4ygFPLA807XPpTQLNGAdIX8 eWj3TWsZGZVBUXVDS60bmTXiCGtkOype3bNIlr/YvfTIhMKKMowyud9QEbKtD1f/qqMlwJ m1emELrWIhgoX1qchSoTv8FVCbYonauOnVc/h2ij80WPh1eCfFVqD7tg7oj0WMYRC0rnaH 6KoB2UvAi3xLmzTMt1u1MJdqC1v4YPvYYQiyl6cbvP8aQdTIr2O1Bc0GX7Cg9A==
X-CM-Envelope: MS4xfBqYl+aEB1VA6ALtw4r+jAYqZvXopJzAHlZvSD2YY1f8XiRakAamEx8kMGPYYC0kr88M1g7qGEn//KGbnMmAN0LjSXz+fGfvcUjiyujOKVGxG2UuEVQ8 lU2RPYC6aRvdJtbWWHztzvx00Rb75BSX9BatQNZ/cGuBuNO/qdhmUWBcrn/d2Ly/1wIgDkTcLO/2kGGIAvwXlO3NYBelDrODfJHHvEUP+h2JQImREBGRLiPn NXsBzlnd6ZUBQSXCxH75VUd8hezGurb6lXOP4qDngbqi03aNWMg7K2rL56ubq0PfD2oUlAdiBprbks/iw74IN3NvAoOnQKz1zt/9AL0ow87qJ4H9BG8ifb4X txLKQp4xzNoDqxf9Jhcvn63YIgIcHV6/zX+N2LVf1ft5uTn1F3/Ee8HCeq17URH/FB4mPbAx
X-Soverin-Id: 01a15e21-7770-78bd-9e6a-3cd036249f98
Content-Type: multipart/alternative; boundary="------------hfoO06TucBlmWp44WgdgceUL"
Message-ID: <b80de824-153d-4819-8980-a12e7dff8af9@iotconsultancy.nl>
Date: Wed, 21 Oct 2026 13:03:33 +0200
MIME-Version: 1.0
To: "snac@ietf.org" <snac@ietf.org>
References: <5922305083fc4b799b6ae5551fa2ba93@huawei.com> <SA3PR11MB347061C35178E7F76276B5D1EBA99C2@SA3PR11MB347061.namprd11.prod.outlook.com> <df56f418-c013-4e0f-999a-03e2dd066ea4@iotconsultancy.nl> <7b8ab89b-9bec-4c9f-a31e-245e1d34c21e@app.fastmail.com> <4e64a55a-bc38-43d6-a132-46ab61e8c4f5@iotconsultancy.nl> <CAGwZUDsyBHoD-7pH=ni9=6EqzDnTXCsh9tkWaQVZrOXVqCsswQ@mail.gmail.com> <7225f6dc-ec3f-426a-a085-f7914b45e10b@iotconsultancy.nl> <CAGwZUDszYOhiE784z3sjf_VL48h970zOkUWJdz_1A3_fBiiXbw@mail.gmail.com> <2f3c853b-f74f-412a-8d57-92538eb530ce@iotconsultancy.nl>
Content-Language: en-US
From: Esko Dijk <esko.dijk@iotconsultancy.nl>
Organization: IoTconsultancy.nl
In-Reply-To: <2f3c853b-f74f-412a-8d57-92538eb530ce@iotconsultancy.nl>
X-Spampanel-Class: ham
X-Spamd-Bar: --------------------------
Message-ID-Hash: OI66XW4G3CXZPKJMU2JUNFGSL76AYXG5
X-Message-ID-Hash: OI66XW4G3CXZPKJMU2JUNFGSL76AYXG5
X-MailFrom: esko.dijk@iotconsultancy.nl
X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; loop; banned-address; emergency; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header
CC: Ted Lemon <mellon@fugue.com>, "Eric Vyncke (evyncke)" <evyncke@cisco.com>, Zhe Lou <zhe.lou@huawei.com>, Jonathan Hui <jonhui@google.com>
X-Mailman-Version: 3.3.10
Precedence: list
Subject: [Snac] Re: 2nd WGLC: draft-ietf-snac-simple (ends 2026-09-23)
List-Id: "Mailing list for discussing problems relating to the automatic connection of stub networks to existing infrastructure networks." <snac.ietf.org>
Archived-At: <https://mailarchive.ietf.org/arch/msg/snac/LGaFUn70rbChChU0INx1Yt6BR2k>
List-Archive: <https://mailarchive.ietf.org/arch/browse/snac>
List-Help: <mailto:snac-request@ietf.org?subject=help>
List-Owner: <mailto:snac-owner@ietf.org>
List-Post: <mailto:snac@ietf.org>
List-Subscribe: <mailto:snac-join@ietf.org>
List-Unsubscribe: <mailto:snac-leave@ietf.org>

Hi all,

the new revision -13 is now posted. It contains the fixes/suggestions 
done after WGLC, plus some bugfixes and rewordings.

Notably there's a normative ref added for RDNSS Option (RFC 8106), and 
the use of the All-Nodes multicast address is now explained for being 
able to receive RAs (before it said NS - which aren't sent to that address.)

And the NS probing procedure is now clarified per the recent discussions 
on the list and in GitHub.

@Eric it should be now ready to proceed as you hoped for last week! :-)

Esko


On 10/20/26 13:00, Esko Dijk wrote:
>
> Great, I've now updated the text again based on your comments.
>
> A change I made that's relevant also for the wider WG: the monitoring 
> of routers (via what's now called the Monitored Routers List) is now 
> defined to happen in all states. Before, it was implied but not 
> clearly stated.
>
> For example: when in state UNKNOWN, the discovery of one or more 
> suitable router(s) will add an entry for each in the List right there. 
> Then it transitions to state SUITABLE where that existing List entry 
> keeps being monitored.  Similar for other states. If the List was not 
> kept in all states, things wouldn't work as described.
>
>
> Let me know if any further comments; I plan to submit tomorrow 
> (Wednesday) with this new text.
>
> Esko
>
>
> On 10/17/26 18:09, Jonathan Hui wrote:
>> Hi Esko,
>>
>> Thanks for putting together PR #224. Removing the entry from the 
>> Monitored Routers List when NS probing fails (and letting a 
>> subsequent suitable RA re-add it) is cleaner and avoids needing 
>> separate Active/Unreachable state markings.
>>
>> I left a few minor comments and suggestions directly on PR #224.
>>
>> Also +1 on standardizing on "moves to STATE-..." throughout the 
>> document for consistency and readability.
>>
>> Regarding OpenThread BR SLLAO, I'll submit a couple of PRs to close 
>> that implementation gap.
>>
>> Thanks.
>>
>> --
>> Jonathan Hui
>>
>>
>> On Sat, Oct 17, 2026 at 8:27 AM Esko Dijk 
>> <esko.dijk@iotconsultancy.nl> wrote:
>>
>>     Jonathan, Ted,
>>
>>     I've now added a new PR:
>>     https://github.com/ietf-wg-snac/draft-ietf-snac-simple/pull/224
>>
>>     This is like the previous text but without the usage of NUD: it
>>     now uses unicast NS messages to probe. Also, when NS probing
>>     fails, the router is considered unreachable and the entry is
>>     removed from the monitored routers list. This has the same effect
>>     as "stopping the timer" in that case. It avoids that a
>>     disappeared router is being probed up to 10 mins which would have
>>     the unwanted side effect of multicast NS messages attempting
>>     link-layer address lookup. Now, receiving an RA again from the
>>     same router after it was unreachable for a while treats it just
>>     like a new suitable router coming up. It resets the timer to
>>     ReachableTime (which is/was already randomized).
>>
>>     The new PR proposal is close to the text we already had in -12,
>>     but now more explicit and with the randomization per SNAC router.
>>
>>     (One side note on the text: it removes the entry markings
>>     'Active' / 'Unreachable' which aren't needed anymore in this
>>     solution. If we do want passive monitoring of an unreachable SNAC
>>     router i.e. meaning that getting a random unicast packet from its
>>     address proves that it is still alive, then we would need these
>>     markings. But I assume we don't want that. Current solution is
>>     that only an RA can 'revive' an unreachable router to a suitable
>>     router.)
>>
>>     On OpenThread BR SLLAO: good to know that some implementations
>>     already include it; I've only checked the 'ot-br-posix' project
>>     build so far.
>>
>>     Finally, one minor nit I saw in the current text: we use the
>>     phrases "moves to STATE_SOMETHING" and "moves to the
>>     STATE_SOMETHING state" and "moves to the state STATE_SOMETHING"
>>     interchangably. For brevity and ease of reading I would change
>>     these all to the first form, if there are no objections to that.
>>
>>     regards
>>
>>     Esko
>>
>>
>>     On 10/16/26 06:03, Jonathan Hui wrote:
>>>     Hi Esko,
>>>
>>>     Agreed with Ted on decoupling this from NUD and just using
>>>     unicast NS probes (up to MAX_UNICAST_SOLICIT) with the
>>>     randomized ReachableTime.
>>>
>>>     Along the same lines of avoiding unnecessary probes, I think we
>>>     should also stop the timer once a probe fails (marking the
>>>     router Unreachable until we hear a new RA from it) rather than
>>>     continuing to re-probe every ReachableTime for 10 minutes.
>>>     Similarly, receiving a suitable RA should reset the entry's
>>>     timer to a new randomized ReachableTime. Since ReachableTime is
>>>     randomized between 60 and 90 seconds, that still spreads out
>>>     probes across SNAC routers while avoiding probing right after an RA.
>>>
>>>     On your note about OpenThread BRs and SLLAO: OpenThread's
>>>     RoutingManager actually does include SLLAO in RAs (and RS/NS
>>>     <https://rs.corp.google.com/NS>) if the platform implements the
>>>     link-layer address callback. Platforms like ESP-IDF and Zephyr
>>>     already do this; ot-br-posix just hasn't hooked up that callback
>>>     yet, which is easy to add. Since RFC 4861 already has a SHOULD
>>>     for SLLAO, I don't think we need to add a MUST in this draft.
>>>
>>>     Thanks.
>>>
>>>     --
>>>     Jonathan Hui
>>>
>>>
>>>     On Thu, Oct 15, 2026 at 8:57 PM Esko Dijk
>>>     <esko.dijk@iotconsultancy.nl> wrote:
>>>
>>>         Good point, it wasn't intended to increase multicast usage.
>>>         That can be easily fixed by updating the PR to use only
>>>         unicast NS up to MAX_UNICAST_SOLICIT times and decouple it
>>>         from NUD - so it was an unintended side consequence and not
>>>         a good idea maybe after all. I can send an update tomorrow.
>>>
>>>         Thinking about avoiding multicasts: maybe important to note
>>>         that OpenThread BRs will not include SLLAO in their RAs. So
>>>         on a link with multiple OpenThread BR SNAC routers, that
>>>         will cause quite some multicasts for the first NS probe
>>>         that's done after a multicast RA, because all routers need
>>>         to resolve the link-layer address of the RA-sender using
>>>         solicited-node NS multicast.
>>>
>>>         The proposed PR change happens to reduce the pain of these
>>>         multicasts by introducing the randomized ReachableTime, so
>>>         that at least all these multicasts don't happen at the exact
>>>         same time. A further remedy could be requiring SLLAO in RAs
>>>         and in NS messages ("MUST"), though per today's RFC 4861
>>>         that is already a SHOULD.
>>>
>>>         Esko
>>>
>>>         On 10/15/26 22:15, Ted Lemon wrote:
>>>>         FWIW, I think switching to using NUD here makes the
>>>>         multicast usage of the protocol substantially worse, and I
>>>>         do not support this change. NUD is solving a different
>>>>         problem.
>>>>
>>>>         On Thu, Oct 15, 2026, at 2:44 PM, Esko Dijk wrote:
>>>>>
>>>>>         Hi Eric,
>>>>>
>>>>>         Thanks for reminding. There's only 1 PR (#223)
>>>>>         <https://github.com/ietf-wg-snac/draft-ietf-snac-simple/pull/223>
>>>>>         left to be reviewed and then merged. And 1 open issue
>>>>>         (#118) which I think we can close or leave as future work:
>>>>>         the documents it points to are personal expired drafts.
>>>>>         Rest of the items is left for future work.
>>>>>
>>>>>         I can merge the PR and publish the update by mid next
>>>>>         week:  that gives me some time for a quick implementation
>>>>>         check on the #223 proposed mechanism, and co-authors some
>>>>>         time to check if they're ok with the new text.
>>>>>
>>>>>         regards
>>>>>
>>>>>         Esko
>>>>>
>>>>>
>>>>>         On 10/15/26 18:04, Eric Vyncke (evyncke) wrote:
>>>>>>         Dear authors and SNAC WG,
>>>>>>
>>>>>>         I think that it is great time to merge all PR for the
>>>>>>         authors, formally close this 2nd WG LC, update the
>>>>>>         shepherd write-up (if required) and for the chairs to
>>>>>>         request IESG publication.
>>>>>>
>>>>>>         Obviously, I can be wrong about the state of the document,
>>>>>>
>>>>>>         Regards
>>>>>>
>>>>>>         -éric
>>>>>>
>>>>>>
>>>>>>         *From: *Zhe Lou <zhe.lou=40huawei.com@dmarc.ietf.org>
>>>>>>         <mailto:zhe.lou=40huawei.com@dmarc.ietf.org>
>>>>>>         *Date: *Thursday, 3 September 2026 at 14:47
>>>>>>         *To: *snac@ietf.org <snac@ietf.org> <mailto:snac@ietf.org>
>>>>>>         *Cc: *ipv6@ietf.org <ipv6@ietf.org> <mailto:ipv6@ietf.org>
>>>>>>         *Subject: *[Snac] 2nd WGLC: draft-ietf-snac-simple (ends
>>>>>>         2026-09-23)
>>>>>>
>>>>>>         As the SNAC simple contains substantial changes from the
>>>>>>         past weeks, we would like to start the 2^nd 3-week
>>>>>>         Working Group Last Call for:
>>>>>>
>>>>>>           draft-ietf-snac-simple-12 Automatically Connecting Stub
>>>>>>         Networks to Unmanaged Infrastructure
>>>>>>
>>>>>>         Please review the document and send comments to
>>>>>>         _snac@ietf.org_ <mailto:snac@ietf.org> by 2026-09-23.
>>>>>>
>>>>>>         The chairs are especially interested in whether the
>>>>>>         document is ready to be sent to the IESG for publication
>>>>>>         as a Proposed Standard, and whether there are remaining
>>>>>>         technical or editorial issues that must be resolved
>>>>>>         before publication.
>>>>>>
>>>>>>         Authors and listed contributors: please reply to confirm
>>>>>>         that any IPR of which you are aware has been disclosed in
>>>>>>         conformance with IETF IPR rules.
>>>>>>
>>>>>>         Document:
>>>>>>         _https://datatracker.ietf.org/doc/draft-ietf-snac-simple/_
>>>>>>         <https://datatracker.ietf.org/doc/draft-ietf-snac-simple/>
>>>>>>
>>>>>>         Thanks,
>>>>>>         Darren and David
>>>>>>         SNAC WG chairs
>>>>>>
>>>>>>
>>>>>>         _______________________________________________
>>>>>>         Snac mailing list --snac@ietf.org
>>>>>>         To unsubscribe send an email tosnac-leave@ietf.org
>>>>>
>>>>
>>>>         _______________________________________________
>>>>         Snac mailing list --snac@ietf.org
>>>>         To unsubscribe send an email tosnac-leave@ietf.org
>>>
>>>
>>>     _______________________________________________
>>>     Snac mailing list --snac@ietf.org
>>>     To unsubscribe send an email tosnac-leave@ietf.org
>>     -- 
>>     *IoTconsultancy.nl* | Email/Teams: esko.dijk@iotconsultancy.nl |
>>     +31 6 2385 8339
>>
>>
>> _______________________________________________
>> Snac mailing list --snac@ietf.org
>> To unsubscribe send an email tosnac-leave@ietf.org
> -- 
> *IoTconsultancy.nl* | Email/Teams: esko.dijk@iotconsultancy.nl | +31 6 
> 2385 8339
>
> _______________________________________________
> Snac mailing list --snac@ietf.org
> To unsubscribe send an email tosnac-leave@ietf.org
-- 
*IoTconsultancy.nl* | Email/Teams: esko.dijk@iotconsultancy.nl | +31 6 
2385 8339