Re: [Idr] Request to adopt draft-heitz-idr-large-community - Working Group Adoption call (9/6 to 9/20)

[Jared Mauch <[e-mail]>]

> On Sep 22, 2016, at 12:16 PM, Simpson, Adam 1. (Nokia - CA) <[e-mail]> wrote:
> 
> OK, so the translation should be flexible. Two follow-on questions:
> 
> 1. Is there a need for reverse mapping, from large to 16:16 (RFC 1997)?

I would argue this should be done in the same way, with explicitly configured operator policy.

If I map 2914:9999 to no-export or local-as (for example) I could similarly map 2914:0:9999 to something else with a similar mapping ecosystem.

What I’ve seen (eg: I’ll keep using XR as the example) is take something and do ($1:$2) to tokenize it then I can use those variables within that sub-part of the policy.

> 2. Is it sufficient for mapping rules to be session level only? I.e. towards peer X always replace $x:$y with $x:$y:0. Or are there cases that need more complex policy logic? I.e. Only in BGP routes that match a particular prefix-list/route-filter then replace $x:$y with $x:$y:0.

I’d say if the policy is attached to a session, sure.

One place I might decide to do this is on an iBGP session or eBGP session on import/export.  Obviously anything that passes import would be propagated further in my iBGP or towards RR.

- Jared

> 
> -Adam
> 
> 
> 
> On 2016-09-22, 10:37 AM, "Idr on behalf of Jared Mauch" <[e-mail] on behalf of [e-mail]> wrote:
> 
>> 
>>> On Sep 21, 2016, at 7:39 PM, Jakob Heitz (jheitz) <[e-mail]> wrote:
>>> 
>>>> I will be asking my friendly vendors to have a mapping option to map 1997 into -large-, so I can take my ($1,$2) and
>>>> turn it to $1:$2:0 (or any other order we define) internally.
>>> 
>>> Should we specify rules for such mappings in the draft?
>> 
>> No, because these are opaque.  I may want $x:$y mapped to $x:$y:0 or 0:$y:$x or any of the other variants.  My needs may be different than Gert or another operator.  Giving flexibility in how one self-organizes the bitspace is a feature vs having to strictly constrain the results and ordering.
>> 
>> - Jared
>> _______________________________________________
>> Idr mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/idr