Hi Carles,

Thank you for the detailed answers, and no worries about the delay.

I have gone through -14, and the explicit field sizes together with the new ECN field resolve the points I raised.

Best regards,
Omar

On 31/10/2026 10:54, Carles Gomez Montenegro wrote:
Hi Omar,

Sorry for the late response, and many thanks for your comments!

Indeed, a few details about the PRO Header were not sufficiently specified in the previous version of the draft.

Please find below my inline responses:

On Thu, 22 Oct 2026 at 17:39, omar merroun <[email protected]> wrote:

    Hi all,

    I am Omar, a student at IMT Atlantique. I am working on a school
    project, with two fellow students and under the supervision of
    Laurent Toutain, to implement a SCHC-Lo network using the PRO
    approach.

    Thank you for this work! While reading the PRO Header section
    (Section 4.3 of draft-ietf-6lo-schc-15dot4-13), I came across a few
    points I could not resolve on my own from the text :

    1. Field sizes in the PRO Header: Figure 23 seems to show a 7-bit
       Bit Pointer and a 5-bit Address Residue Length, but I could not
       find these sizes in the text. Could you confirm them?


Yes, they are confirmed. In -14, we have explicitly added the field sizes as well in the text.

    7 bits would
       seem sufficient for the Bit Pointer, since it only needs to skip
       the RuleID (RECOMMENDED 1-16 bits from Section 4.1.3) and any
       Version residue.


Note that, per RFC 8724 (section 10.1) the IPv6 version is not sent (as we are only considering IPv6, not IPv4).

There could be other compression residues as well, but if the only ones are the Hop Limit and the Destination Address residues, the Bit Pointer indeed only needs to skip the RuleID.


    2. Fields between Traffic Class and Hop Limit: Section 4.3 says the
       Hop Limit and destination address residues follow the Traffic
       Class residue, and Section 6.1 keeps the other fields in IPv6
       header order. This seems to imply that Flow Label, Payload Length
       and Next Header must have zero-length residues (not-sent or
       compute-*). Is this a requirement for PRO Rules ? If not,
       how does a 6LR locate the Hop Limit residue when the Flow Label
       is sent for example?

    3. Traffic Class residue size: the HLM flag gives the Hop Limit
       residue size and the Address Residue Length gives the destination
       address residue size, but I found nothing indicating the size of
       the Traffic Class residue. Since a router may read and modify the
       ECN bits, should the Traffic Class residue always contain at least
       the 2 ECN bits (MSB(6) with LSB, or the full 8 bits) ? and what is
       expected if Traffic Class is not-sent? Relatedly, Section 3.5.3
       only mentions the Hop Limit and destination address residues as
       located by the pointer; so should Traffic Class be added there?


We addressed both your comments 2 and 3 by modifying the PRO Header format.

As shown in -14, we have added an ECN field to the PRO Header format, which allows a 6LR to read and modify (if needed) the ECN bits of the compressed packet without the need to store C/D Rules:
        0 1 2 3 4 5 6 7 0 1 2 3 4 .... 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3
       +---------------+-+ - - - +----+-------------+---+-----------+
       |      SCHC     |C|       |    |             | E |  Adress   |
       |     Pointer   |I|  DCI  |Ctrl| Bit Pointer | C |  Residue  |
       |    Dispatch   |D|       | Hdr|             | N |  Length   |
       +---------------+-+ - - - +----+-------------+---+-----------+
The rest of the Traffic Class field can be handled like the other IPv6 header fields (except for Hop Limit and Destination Address, which have a special handling in PRO).

We hope that the updates and responses above will address your comments!

Cheers,

Carles

    Best regards,
    Omar
_______________________________________________
6lo mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to