Dear all,

I have updated the shepherd writeup as per Éric's points below.

Cheers,

Carles (as document's shepherd)

On Wed, 8 Jul 2026 at 16:03, Eric Vyncke (evyncke) <[email protected]>
wrote:

> # Éric Vyncke, INT AD, AD review for draft-ietf-6lo-nd-gaao-10
> CC @evyncke
> Thank you for the work put into this document. Please find below my AD
> review.
>
> As the responsible AD, I expect all the points below to be addressed,
> either by a revised I-D, or an email reply. Of course, authors and WG can
> reject my points, but this needs to be justified. Once all the points are
> addressed, I will proceed with the publication process, i.e., IETF Last
> Call with a specific copy to 6MAN.
>
> Special thanks to Carles Gomez for the shepherd's detailed write-up
> including the WG consensus and the justification of the intended status
> *BUT* see below.
>
> I hope that this review helps to improve the document,
>
> Regards,
>
> -éric
>
> Note: this AD reviews follows the Markdown syntax of
> https://github.com/mnot/ietf-comments/tree/main, i.e., they can be
> processed by a tool to create github issues.
>
> ## Critical issues
>
> ### Update to Shepherd's write up
>
> The write-up must be refreshed about the DHCP discussion at IETF-125 and
> the new appendix A.
>
> About Q11, I suggest adding some words about "interoperation" as this is
> critical for a PS.
>
> ### Canonical IPv6 addresses
>
> Ensure that *all* IPv6 addresses follow RFC 5952 (notably in lower case,
> max compression).
>
> ### IANA TBD
>
> As there are more than one "TBD" value, please use "TBD1", "TBD2", ...
>
> ### Bytes vs. octets
>
> Some people prefer using "octets" rather than "bytes".
>
> ### Address privacy
>
> Should there be some recommendations for AAF specifications to state their
> status about privacy of addresses ? This could be a paragraph in the
> security section.
>
> ### Abstract
>
> Add a line about why DHCPv6 is not optimum for LLN.
>
> ### Section 2.2
>
> Also list the RFC for IID, NS, NA, ...
>
> ### Section 2.3
>
> Please specify whether GAAO is a AAF.
>
> ### Section 3
>
> `The IPv6 address assignment model within a local domain relies on
> randomly generated Interface Identifiers (IIDs)` is not correct as it is
> not always `randomly` and it is rather a layer-2 domain than the ambiguous
> `local domain`.
>
> `In the latter case, additional mechanisms are required to ensure
> address` also has some problems:
>
> 1) AFAIK, even using DHCPv6 the nodes are required to do DAD
> 2) CGA & SEND are mostly history :-( and were never required.
>
> In figure 1, unsure whether the top left "STEP" is useful, please consider
> removing it.
>
> s/is required and proposed/is required and *defined*/ as this is a PS
>
> ### Section 4
>
> As the R-bit is not set, it is explicit and no `it *implicitly* indicates`
>
> About AAF and `it is RECOMMENDED` why not a MUST ? See
> https://datatracker.ietf.org/doc/statement-iesg-statement-on-clarifying-the-use-of-bcp-14-key-words/
> for the required guidance when the recommandation is not followed.
>
> A verb/word is probably missing in `However, a 6LN MAY use a value
> different from 0, meaning that it is requested to use a specific known AAF`
>
> In `Address/prefix lifetime SHOULD be configurable` on the requester or on
> the 6LR ? Also missing the required guidance per above IESG statement.
>
> ### Section 5
>
> `When a node bootstraps`Is this the only case ? I.e., no expiration
> renewal ?
>
> ### Section 5.1
>
> Use a "MUST" in `the AAF field SHOULD be set to zero unless by
> configuration ` as there is a "unless".
>
> In step 1), specify the source & destination IPv6 addresses.
>
> In step 2), should "on the link" be added after `If the node is a 6LR it
> MUST use the same AAF to generate addresses/prefixes to requesting neighbor
> nodes` (`6LoWPAN network` is also used in other places in the document,
> please ensure to use consistent wording for the same concept)
>
> ### Section 5.2
>
> `The use of this option is done by configuration. ` on which node (6LN
> and/or 6LR) ?
>
> ### Section 5.3
>
> Please specify that the optimisation must be done within the MTU limit as
> NDP messages cannot be fragmented.
>
> Suggest to swap 5.3.1 and 5.3.2 to respect the chronology.
>
> ### Section 5.4
>
> s/SHOULD/MUST/ in `When the node receives this status back it SHOULD
> perform one of the following actions` as all cases are covered.
>
> ### Section 6
>
> Specify RFC 7400 *section 3.3*.
>
> Out of curiosity, is there any reason why the M-flag is in the 2nd 32-bit
> rather than in the first 32-bit ?
>
> ### Section 7
>
> `This section provides guidance to ` it is not about "guidance" but about
> "request", so, be clear.
>
> In all subsection, add informational references to the IANA registries,
> e.g.,
> https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml#icmpv6-parameters-5
>
> The M-bit position is specified in section 6 so do not use TBD (see my
> other comment about "TBD")
>
> ### Section 7.4
>
> Should "6LowPan" be added in the registry name of `"Generic Address
> Assignment Option".` ?
>
> The parent registry should also be named.
>
> Do we really want to use FCFS in a registry with only 14 available code
> points ? Specification required or even standard action seem to be
> preferrable.
>
> ### Section 8
>
> Why not "MUST" rather than "SHOULD" ? See also the IESG requirement above.
>
> ### Appendix A
>
> First, thanks for adding this useful appendix.
>
> I do not think that `While DHCPv6 is the standard ` is correct as most
> networks can only rely on SLAAC as some widespread OS do not support
> DHCPv6-IA... especially as later in the text there is `for actual Address
> Assignment, it is almost non-existent in the field`. So, let's be
> consistent ;-)
>
> ### Section A.1
>
> Suggest adding whether is it multicast or unicast.
>
> ### Section A.3
>
> s/address several, messages need*s*/address*,* several messages need/
>
> More important, in DHCPv6 there are at least 2 multicast messages: SOLICIT
> & REQUEST, the other 2 can be unicast (rarely and obsoleted by RFC 9915) or
> multicast.
>
> The appendix should include DHCPv6 rapid commit as it is only 2 messages.
>
> Should RFC 9119 also be a reference in the appendix ?
>
> ## Non-critical / cosmetic issues
>
> Note: these points must also be addressed.
>
> ### Abstract
>
> s/a *new* extension/an extension/
>
> ### Section 4
>
> s/a*n* Neighbor Adverti/a Neighbor Adverti/
>
> Suggest repeating that the presence of `Address/Prefix` is solely based on
> the GAAO length header field.
>
> ### Section 5.2
>
> s/the 6LR is expect to/the 6LR is expect*ed* to/
>
> ### Acknowlegments
>
> s/for helping clarifying/for helping *to clarify*/
>
> ### Section A.2
>
> s/GAAO does not need to be tramistted /GAAO does not need to be
> *transmitted*/
>
> s/piggy*-*backed/piggybacked/ (and other places)
>
> ### Section A.3
>
> s/specific packets need*s*/specific packets need/
>
> I.e., suggest to use a spell-checker also on the new sections, e.g.,
> `reesources`, `consists on a two unicast messages`, `likelyhood`, `twise`
>
> ### Use of SVG graphics
>
> To make a much nicer HTML rendering, suggest using the aasvg too to
> generate SVG graphics. It is worth a try especially as this I-D uses the
> Kramdown file format by starting the block with `~~~ aasvg` ;-)
>
> ### References by RFC Number
>
> I must admit that I am very bad with RFC numbers so when reading
> `procedures defined in [RFC8505], [RFC9685], or [RFC9926]`, it means little
> to me :-( when possible try to use a short name (in addition to the RFC
> numbers such as SLAAC, IPv6, DHCPv6, ....)
>
>
>
_______________________________________________
6lo mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to