Brian E Carpenter wrote:
Bob,
On 2009-04-03 14:18, Robert Moskowitz wrote:
<snip>
Can we develop a list of techincal issues that are drivers for
IPv6NATs?
Address Independence
and multihoming?
Topology hiding
Address Amplification
Is this tied into prefix allocations or is this a separate one?
And what else?
Then we define what each means and offer solutions today for them.
See RFC 4864.
Would that be only sec 6?
That was the intention of the authors of 4864, but we might have missed
something.
However, address amplification is *not* on my list. We really need ISPs
to give (well, actually rent out) short enough prefixes that this
excuse for NAT utterly vanishes.
Well if you're goal is to design NATs out of the IPv6 deployments and
you recognize the root driver for address amplification, then your goal
is to remove that driver so there will not be a need for address
amplification driving a NAT market.
Did I express that well enough? Remember, I look at my wording through
my dyslexic eyes....
Anyway, so we have to ensure that address prefixs are 'right sized'.
Does this mean nothing larger than /62? /60? That we issue a BCP to
boycott any ISP issuing /64s or single addresses (hotel rooms?)?
From what I am hearing, Address amplification is a problem that drives
NAT and a problem that can 'easily' be mitigated.
_______________________________________________
nat66 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nat66
