Erik Romijn wrote:
Hi all,

I'm working with Juergen Kammer to make Quagga's MRT dump code and
libbgpdump support 32bit ASNs. We have a few proposals for things that
appear to be missing in draft-ietf-grow-mrt-03.

We've encountered a few things in the library which are supposed to be
based on working group discussions, but are not included in the
(expired) draft.
However, the old archives are inaccessible and have not been merged with
the new ones. So, I might be repeating a bit of previous discussion here.

What we propose to be included in the draft:

- A new MRT routing information type: TABLE_DUMP_32BIT_AS.
  A current implementation in the library is to encode the 32bit AS flag
  into the address family field of the TABLE_DUMP, but this doesn't seem
  very nice.
  TABLE_DUMP_32BIT_AS would be the same as TABLE_DUMP, except with a
  32bit field for the Peer AS.

      Eric,
         My assumption is that implementations would move to
using the BGP4MP_ENTRY type for table dumps as it was meant
to replace the TABLE_DUMP type.    The BGP4MP_ENTRY header
does not encode an AS number in the header and is transparent to
32-bit AS numbers.   I guess I could either create a new
TABLE_DUMP_32BIT_AS type or add text referring to  the
BGP4MP_ENTRY type for 32-bit AS support.

- A new BGP4MP type: BGP4MP_STATE_CHANGE_32BIT_AS.
  This would be the same as BGP4MP_STATE_CHANGE, except with a 32bit
  field for the source and destination AS number.

   Okay, I had thought that RFC4271 had changed the states in
the state machine and thus would result in the deprecation of the
BGP4MP_STATE_CHANGE type, but I see it has the same
6 states as part of it's FSM.   I can go ahead and add this type.

 -Larry

Any comments are welcome.

cheers,
------------------------------------------------------------------------

_______________________________________________
GROW mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/grow


!DSPAM:45ab6d5c13361242917256!


_______________________________________________
GROW mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/grow

Reply via email to