ietf
[Top] [All Lists]

Fwd: myth of the great transition

2003-06-20 11:17:22


Begin forwarded message:

From: Keith Moore <moore(_at_)cs(_dot_)utk(_dot_)edu>
Date: Thu Jun 19, 2003  8:43:18  PM America/Montreal
To: S Woodside <sbwoodside(_at_)yahoo(_dot_)com>
Cc: moore(_at_)cs(_dot_)utk(_dot_)edu
Subject: Re: myth of the great transition

OK, so let's say I'm the author of a voice over IP application on a
platform that supports IPv6, like, say, Mac OS X. What part of IPv6 is
going to make this work. I would need NAPT-replacement box to support
IPv6 as well, right? It would assign an IPv6 address to each of the
inner systems, and then ... what? I guess my caller who's trying to
phone me, I give her my IPv6 address. She connects to that address, and
her magic box sets up an IPv6 tunnel to my magic box automatically.
Then my magic box forwards the packets to the right machine in my
network.

Add a firewall to that, and you've got something that replaces NAPT.

I'm thinking maybe you could keep IPv4 NAPT as a legacy feature for
inside hosts and applications that don't support IPv6 yet. But apps
that do support IPv6, would not have to do any work to traverse the
NAPT.

you've just described my router at home, I think. except that there's really no overhead in setting up a tunnel to your magic box - the sending router
just encapsulates the IPv6 packet within an IPv4 packet and sends it.
there's no negotiation or per-tunnel state required.

and yes, it does have NAPT as a legacy feature - I use it for web browsing and
reading email.



--
www.simonwoodside.com -- 99% Devil, 1% Angel




<Prev in Thread] Current Thread [Next in Thread>