I sent this to "fun" back in July, it bounced, but I didn't get onto
homenet until now.  I will read archives now.

Date: Wed, 27 Jul 2011 13:15:33 -0400
Sender: [email protected]


Section 3.2 is the meat, and I think maybe we will violently agree on
many points, and mis-understand our objections on others.

Let me start then with:  Largest Possible Subnets

I think that this is an interesting assertion.  

I find that there has been an over reliance on too-extensive layer-2
stuff, including trying to bridge wired and wireless mechanisms is a
result of: 
       - scarcity of IPv4 networks
       - combined with lack of deployed multicast (and/or too much
         reliance on broadcast) between subnets.

I don't think that IPv6 should continue this "tradition".
I believe that L2-bridging should be the exception rather than, 
as this suggests, the rule.

For instance, I would far prefer that my laptop got a prefix (via DHCPv6
PD), and used that for the link over USB (or PAN) to my smartphone.
I really don't want it to bridge.   My smartphone might (or might not)
also have a link on the same wifi, or maybe not. or maybe it has IPv6
via 3G...  
If everything is bridged, then there are lots of loops possible, while
we have STP, I think it's a fail.
(I could go on with many more use cases, I suspect that we should)

Or, I think that figure 3 is far more likely than figure 1, even in the
face of no explicit security requirements to seperate things.

-- 
]       He who is tired of Weird Al is tired of life!           |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] [email protected] http://www.sandelman.ottawa.on.ca/ |device driver[
   Kyoto Plus: watch the video <http://www.youtube.com/watch?v=kzx1ycLXQSE>
                       then sign the petition. 

Attachment: pgpfUtLWOQjOc.pgp
Description: PGP signature

_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to