Re: Ordering of NHRP extensions

[Grenville Armitage <[e-mail]>]

>>>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
>>
>>I hope not.  This might in a few cases speed up receive processing
>>slightly, at the cost of a more substantial hit in time and complexity
>>at the transmit end.

Could you please justify this assertion. The 'hit' you get by forcing
programmers to structure their TX code cleanly is not a good reason.
Since the TX side knows what it wants, the set of options to be TXed
can be built pre-transmission to be in the most optimal order for RX
processing. Minimal per-packet TX side hit.

	[..]
>>This sort of idea has been floated in other places and generally
>>rejected.  I think it should be rejected here as well.

Not thinking of CIF recently, are we? :)  Perhaps if you could
explain why it was rejected in the past, the WG would be in a better
place to understand how the past lessons ought to be applied here.

I happen to think its a reasonable idea. RX processing may
be improved (and could hardly be made worse), and I dont see
where the hit is on TX processing.

cheers,
gja
_________________________________________________________________________
Grenville Armitage                               [e-mail]
Bellcore, 445 South St.      http://gump.bellcore.com:8000/~gja/home.html
Morristown, NJ 07960 USA          (voice) [phone] {.. 2504 (fax)}