Hi Renato,

Thanks for the feedback. Please see my response below inline.

Thanks,
Yingzhen


On Mon, Apr 21, 2025 at 12:07 PM Renato Westphal <[email protected]>
wrote:

> Hi Yingzhen and Acee,
>
> I'm glad I could be of help, and thanks for incorporating my suggestions :)
>
> I've updated Holo to implement the latest version of the SR-MPLS
> modules. Here are some more examples of JSON data for IS-IS and OSPF:
> * IS-IS: https://textbin.net/raw/jfpbj1axgm
> * OSPFv2: https://textbin.net/raw/ksuukedldr
> * OSPFv3: https://textbin.net/raw/sbj5pgaxf1
>
> [Yingzhen]: these examples actually have LS database. I'm kind of
wondering whether we should include them in an appendix. Thoughts?

While updating the code, I noticed a few minor inconsistencies that I
> thought were worth sharing:
> * In the OSPF module, SIDs are always represented using a
> "label-value"/"index-value" choice. However, in some cases there's an
> additional "length" leaf, and in others, there isn't.
> * In the IS-IS module, SIDs are represented either as a simple "sid"
> leaf or as a "label-value"/"index-value" choice combined with a
> "length" field.
>
[Yingzhen]: Thanks for catching this. I looked at RFC8667 and fixed the
is-is model to be consistent with OSPF.


> * In the IS-IS module, the "adjacency" list is augmented with a list
> of local Adjacency-SIDs ("adjacency-sid"). In the OSPF module, the
> "neighbor" list doesn't have a similar augmentation. Adjacency-SID
> information is only available in the LSDB.
>

[Yingzhen]:
 do you mean the following in IS-IS?

  augment
/rt:routing/rt:control-plane-protocols/rt:control-plane-protocol/isis:isis/isis:interfaces/isis:interface/isis:adjacencies/isis:adjacency:

    +--ro adjacency-sid* []

       +--ro value?                  uint32

       +--ro address-family?         iana-rt-types:address-family

       +--ro weight?                 uint8

       +--ro protection-requested?   boolean


This is an augmentation of operational state. You're right, we don't have
this in OSPF. I looked at RFC9129, this can be added to
ospf:interface/ospf:neighbors/ospf:neighbor. Unfortunately we have
different names in RFC9129 and 9130, neighbors vs. adjacencies.

>
> Best Regards,
> Renato.
>
> Em sex., 18 de abr. de 2025 às 09:15, Acee Lindem
> <[email protected]> escreveu:
> >
> > Hi Renato,
> >
> > Thanks so much for your implementation. All the reviews are also useful
> but this is invaluable.
> >
> > Yingzhen has incorporated your comments in -28.
> >
> > > On Apr 16, 2025, at 8:17 PM, Renato Westphal <[email protected]>
> wrote:
> > >
> > > Hi Acee,
> > >
> > > Thank you for your quick response.
> > >
> > > Em seg., 14 de abr. de 2025 às 12:06, Acee Lindem
> > > <[email protected]> escreveu:
> > >>
> > >> Hi Renato,
> > >>
> > >>> On Apr 13, 2025, at 11:50 AM, Renato Westphal <
> [email protected]> wrote:
> > >>>
> > >>> Hi all,
> > >>>
> > >>> I'm implementing this draft and would like to share some feedback.
> > >>>
> > >>> There's an augmentation to the "router-capabilities" container that
> is
> > >>> wrong. Instead, it is the "router-capability" list within that
> > >>> container that should be augmented.
> > >>
> > >> Mahesh noted this as well and it is fixed in GitHub.
> > >
> > > That might have been a different issue. I just checked the latest -27
> > > version and the augmentation to "router-capabilities" is still
> > > incorrect.
> >
> > Right - that was a different problem with a redundant level of
> hierarchy.  This is in -28.
> >
> > >
> > > As I progress with my SR implementation, I came across another issue
> > > related to augmentations. The module is currently augmenting the
> > > "extended-is-neighbor" and "mt-is-neighbor" containers, instead of
> > > augmenting the "instance" list within those containers. I believe it
> > > would be better to have the Adj-SID sub-TLVs directly associated with
> > > their respective IS reachability entries, rather than grouping all
> > > Adj-SID sub-TLVs for each "neighbor-id" together (in the case of
> > > parallel links, there might be multiple IS reachability entries for
> > > the same neighbor).
> >
> > I agree - this is where all the feature specific neighbor extended
> attributes belong. This is in -28.
> >
> >
> > >
> > > I also suggest renaming the "sid-list" list to "adj-sid-sub-tlv" for
> > > consistency with the existing "prefix-sid-sub-tlv" list and to align
> > > with the OSPF SR module. It might also be a good idea to group all
> > > Adj-SID sub-TLVs inside an "adj-sid-sub-tlvs" container for
> > > consistency as well.
> >
> >
> > Good suggestion - this is in -28.
> >
> > >
> > > Other than these minor issues, I think the module looks excellent.
> > >
> > > FWIW, I'm sharing [1] an example data file populated from a virtual
> > > IS-IS SR topology using the Holo routing stack. It's a partial
> > > implementation of this module, including fixes for the aforementioned
> > > issues.
> > >
> > > [1]
> https://gist.githubusercontent.com/rwestphal/aa85a402e7ed12f6da131b1a0e172c93/raw/dfe4cdbbac4abd21828e5dc9a9a60e5a5d51f9c1/ietf-isis-sr-mpls.json
> >
> >
> > Cool!!!
> >
> > Thanks,
> > Acee
> >
> >
> >
> >
> > >
> > > Best Regards,
> > > --
> > > Renato Westphal
> >
>
>
> --
> Renato Westphal
>
> _______________________________________________
> Lsr mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
>
_______________________________________________
Lsr mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to