> While reviewing draft-ietf-lisp-lcaf-12 I realized that if we want 
> instance-id support for other LCAF types then we must support recursive 
> LCAFs, i.e., use Type 2 and within it store the type we’re interested in. 
> Wouldn’t it be better to have the instance-id value and mask part of the LCAF 
> header? Apart from simplifying processing on xTR side, this would also help 
> avoid having other types, like Type 9, define their own instace-id fields. 

Well Florin, the way I look at this is from this perspective. If we want 
geo-prefix support for all other LCAF types we also have to do nested LCAFs. So 
be all things to all people would make the encoding a combinatorial problem. So 
nesting is really the only way to manage it.

I agree that instance-ID may be a popular LCAF to use but I see others that 
will be just as popular. To name two, AFI-List and RLE.

Dino

> 
> Thanks, 
> Florin
> _______________________________________________
> lisp mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/lisp

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

Reply via email to