I looked at this again and the long Email thread and agree with Peter, Les, and 
Joel that we don't need a protocol specific registry for the Endpoint behaviors 
and associated SIDs. 
Thanks,
Acee

On 3/11/20, 1:42 PM, "Lsr on behalf of Joel M. Halpern" <[email protected] 
on behalf of [email protected]> wrote:

    It does seem to me that using a registry to capture the relationship 
    between the OSPF or IS-IS advertisement (TLV, sub-TLV, ...) and the SR 
    behavior (as defined in the NP registry and subsequent additions) is 
    useful.  I would not want to have to respin the base draft to add 
    additional behaviors.  We use registries for a reason.
    
    Now, if the advertisement registry has a note column, we could use that 
    to record the information.  As far as I know the current registries are 
    not well-structured for such auxiliary information.
    
    Yours,
    Joel
    
    On 3/11/2020 11:52 AM, Christian Hopps wrote:
    > 
    > 
    >> On Mar 11, 2020, at 10:38 AM, Les Ginsberg (ginsberg) 
<[email protected]> wrote:
    >>
    >> Chris -
    >>
    >>
    >>>
    >>> Do you think we should get rid of the "used in" columns in the IS-IS 
TLV and
    >>> sub-TLV registries? The documents that define those TLVs and sub-TLVs
    >>> already indicate which PDUs and TLVs they go in, so why do we need that
    >>> info in the registry?
    >>>
    >> [Les:] The difference for me is that the definition of sub-TLVs 
associated with the related set of TLVs is scattered across multiple RFCs. The 
additional information in the registry allows us to find this information in 
one place.
    >> Here, there is only one relevant IS-IS draft on this technology (SRv6). 
If the set of behaviors which can be advertised in IS-IS changes, then an 
additional IS-IS document (or a bis) will be written - and it likely would be 
required for other reasons.
    >>
    >> We still may not agree - but I hope we at least understand each other 
better.
    > 
    > Is the SR network programming document expected to respin a -bis for 
adding a behavior?
    > 
    > Anyway, unless someone else thinks this is important, I guess we can just 
revisit it when the next SR End behavior gets added and we're faced with 
re-spinning this document, the OSPF companion document, and whatever other 
protocols support documents for advertising the new behavior, in order to 
support it (rather than e.g., it just going directly in a base document or a 
combined protocols document and updating a couple registries).
    > 
    > Thanks,
    > Chris.
    > [as WG member]
    > 
    >>    Les
    >>
    >> _______________________________________________
    >> Lsr mailing list
    >> [email protected]
    >> https://www.ietf.org/mailman/listinfo/lsr
    >>
    > 
    > _______________________________________________
    > Lsr mailing list
    > [email protected]
    > https://www.ietf.org/mailman/listinfo/lsr
    > 
    
    _______________________________________________
    Lsr mailing list
    [email protected]
    https://www.ietf.org/mailman/listinfo/lsr
    

_______________________________________________
Lsr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lsr

Reply via email to