Hi Raul -
On Tue, 19 Apr 2011 17:20:44 -0400, "Raul A. Jager W."
If I understand correctly, a IPv6 only client can reach IPv6 (of course)
and IPv4 by prefixing 80 zero bits to the IPv4 address, But an IPv4
client can not access an IPv6 host. It requires to pass trough some kind
That's what I was thinking. IPv6 addresses are MUCH longer than IPv4
address (128 bits versus 32 bits). So an IPv4 address can be used as
an IPv6 addresses as you noted. (I looked it up on Wikipedia. It's
80 zero bits followed by 16 one bits followed by the IPv4 address. But
not operating systems' IP stacks are programmed to support it.) But
128 bits can't be stuffed into 32 bits even with a shoe horn.
The Wikipedia article talked about tunneling by encapsulating IPv6
packets inside of IPv4 packets with protocol 41 or other methods.
So getting back to David's original question ... I guess you could
keep an internal network on IPv4 indefinitely by using tunneling. But
would you really want to? If the software supports the client side of
IPv6 tunneling over IPv4, then it almost certainly supports IPv6
directly, so why not just use it?
If this is realy like I think, then there is no reason to keep IPv4 in
the internal network, except the cost of the change. My idea is to use
only IPv6 in all new computers, and let the old IPv4 die litle by litle.
That seems reasonable. I'm not responsible for any networking stuff
at work, but in looking at email headers trying to track down why some
emails were being delayed, I noticed that some of the Received:
headers for transfers between our mail servers had IPv6 addresses. So
while I haven't heard anything out of our corporate people about
switching to IPv6, obviously they're doing stuff.
In 1990 the IPv4 addresses got almost exausted, but somedy found a patch
using NAT, and that prolonged the agony of IPv4 until now.
NAT and CIDR both helped delayed the inevitable, but now it's here.
Opinions expressed are my own and do not necessarily represent the views
of my employer or anyone in their right mind.