Re: Ordering of NHRP extensions

["James V. Luciani" <[e-mail]>]

David,
> Is there any implied ordering of extensions in an NHRP packet?
Good idea! :-)  I actually suggested this (via proxy of Andy Malis)
at the last European based IETF.  I explicitly asked for 2 things:
1) the extensions be placed in numerical order by type code
2) (slightly different topic) Next Hop Entries always be
   returned in order from highest to lowest preference with the highest
   in the mandatory part of the NHRP res reply.
Both of these sound eminantly reasonable to me and would make packet processing
faster on the receive side (on the other hand, I have argued many times
that NHRP is NOT fast path :-))

> 
> Is there any requirement for the extensions to be kept in the same 
> order when forwarded?
> I can envisage an implementation where an NHS would always
> send out extensions in some particular fixed order, e.g. have some
> code that looks like :-
> 	if (input_packet.destination_prefix_present)
> 		add_destination_prefix_extension(out_packet)
> 	if (input_packet.NBMA_subnetwork_extension_present)
> 		add_NBMA_subnetwork_extension(out_packet)
> 	etc
> 
> Could the next draft of NHRP add a positive statement on the 
> subject one way or the other?

What do folks think about the above suggestions?