Re: [Idr] I-D Action: draft-ietf-idr-large-community-04.txt

[David Freedman <[e-mail]>]

Apologies for the ultra-late-last-minute input, but I was discussing this with Job offlist and we felt it should be seen by list participants.

My point was about RFC1997 parity and the length validation. 

 " A Large BGP Communities attribute with a length of zero MUST be ignored upon receipt and removed when sending." 

Later wording here has been discussed to be amended (I understand) to:

" A Large BGP Communities attribute with a length of zero MUST NOT be sent and MUST be ignored upon receipt."

Well, I wanted to point out, that RFC1997 was of course updated by RFC7606.

In (7606), the following two comments are made:

" For all path attributes other than those specified as having an attribute length that may be zero, it SHALL be considered a syntax error for the attribute to have a length of zero." 

And 

" The error handling of [RFC1997] is revised as follows: The Community attribute SHALL be considered malformed if its length is not a non-zero multiple of 4" 

So really, if you apply RFC7606 to RFC1997 you have lost parity if you accept a zero length community (even if you just ignore it). 

My suggestion here is just to drop the passive receive tolerance of the zero length largecomm, since it seems out of place (even if it is harmless). 
and let it be considered malformed and handled by RFC7606. 

Thoughts? 

Dave.