Re: [DNA] leaving out options on Landmark RA=Yes(draft-pentland-dna-protocol-00.txt)?

[Greg Daley <greg.daley@eng.monash.edu.au>]

Hi Sathya,

Sathya Narayanan wrote:
 > Greg -
 >
 > Under the assumption that the 'Yes' case is more likely, i.e.
 > the ratio AP/AR is big, it makes sense to emphasize the
 > importance of this optimization by making it a MUST/SHOULD, IMHO.
 > With the size of PIOs, possibly many of them and wireless
 > channels etc., there is possibly significant savings in
 > link-utilization.

I think it is a fairly important issue (since people took
the patience to explain it to me).

So I think a SHOULD (limit the options) is OK.

IMO, MUST is really for
a) Interoperability (as per RFC2119)
b) preventing damage to the network.

The damage is light here particularly for
media like Wireless LAN where the rate of messages
is more important than their size.

I can't see a reason to say MUST in this case.

SHOULD is still a strong statement in IETF.

Greg