Re: [Bier] WGLC: draft-ietf-bier-non-mpls-bift-encoding

Hi Senthil,

Thanks for the comments.

See response inline.

> I believe the ‘Domain-wide unique’ is the key value of Non-MPLS encapsulation.
> Allocation BIFT-id locally on each node is an extra cost, not only in data-plane, but also in control-plane.
>  
> That has been concluded in RFC8296 section 2.2.1.1:
>    In non-MPLS networks, a BIFT-id MUST be assigned for every
>    combination of <SD, SI, BSL> that is to be used in that network.  The
>    correspondence between a BIFT-id and a particular <SD, SI, BSL>
>    triple is unique throughout the BIER domain and is known to all the
>    BFRs in the BIER domain.
>  
> The draft we are discussing is to define the above ‘correspondence’  rule. 

ICE: Agree with Senthil.

>  
> Support !
> However, i would like to get couple of clarifications with respect to the BIFT-id in non-MPLS networks..
>  
> 1. 
>  
> Excerpts from section 2.2.1.1 from RFC8296
>    In an MPLS network, since the BIFT-id is an MPLS label, its value may
>    be changed as a BIER packet goes from BFR to BFR.  In a non-MPLS
>    network, since the BIFT-id is domain-wide unique, it is not expected    // Senthil
>    to change as a BIER packet travels.
>  
> Senthil// I believe, we do not want to mandate that the BIFT-id must be domain-wide-unique in case of non-MPLS networks (for the sake of flexibility).
>               If its *not* domain-wide-unique, then the BIFT-id may change in BFR's as the BIER packet travels from BFIR to BFER.
>               Authors & Fellow BIER'ers, let me know what do you think ?
>  
> 2.
>  
> draft-ietf-bier-non-mpls-bift-encoding-01 proposes two encoding methods for BIFT-id and is mentioned that, both of the encoding options are to derive "domain-wide-unique: BIFT-id's.
>  
> I believe, option-2, that is  "6. The Non-MPLS Static IBU-SI BIFT Encoding", is flexible and can be used for non-domain-wide-unique BIFT-id's also.
>  
> If we agree that, we should allow "non-domain-wide-unique" BIFT-id's for non-MPLS networks, can we re-phrase the text in draft-ietf-bier-non-mpls-bift-encoding-01 to specify that, such an encoding option can be used for non-domain-wide-unique cases as well ?

ICE: The 'IBU-SI BIFT Encoding’ is indeed flexible, but with that the flexibility is in the way you can add additional future arguments since the IBU is like a Label. But, allowing to change the IBU value from hop-by-hop like MPLS is adding a lot of complexity and we’re re-inventing MPLS. So lets keep this simple and domain side unique.

Thx,

Ice.

>  
> Thanks,
> Senthil
>  
>  
> On Sat, Jan 26, 2019 at 1:43 AM Greg Shepherd <gjshep@gmail.com> wrote:
>  
> Change of subject. This is WGLC for:
>  
> https://datatracker.ietf.org/doc/draft-ietf-bier-non-mpls-bift-encoding/
>  
> Same two week timer.
>  
> Thanks
>  
> On Fri, Jan 25, 2019 at 10:13 AM Greg Shepherd <gjshep@gmail.com> wrote:
> Please read and reply to this thread with your vote for/against adoption of:
>  
> https://datatracker.ietf.org/doc/draft-ietf-bier-non-mpls-bift-encoding/
>  
> ..as a BIER WG document.
>  
> This starts a two week counter.
>  
> Thanks,
> Chairs
> (Shep)
> _______________________________________________
> BIER mailing list
> BIER@ietf.org
> https://www.ietf.org/mailman/listinfo/bier
> _______________________________________________
> BIER mailing list
> BIER@ietf.org
> https://www.ietf.org/mailman/listinfo/bier