On Thu, Aug 14, 2008 at 6:19 PM, Marc Manthey <[EMAIL PROTECTED]> wrote:
>> From: Eric Anopolsky <erpo41 at gmail.com>
>> What do you all think about using IPv6 for this purpose?
>
> i need to second that , thats sounds like a great idea ;)
> what would be required ?
> cheers
> Marc
> F.Y.I.

If both hosts have public IPv6 addresses then NAT penetration is not
required. On the server side, cspace needs to publish the host's IPv6
address to the DHT. On the client side, cspace needs to look for and
connect to the IPv6 address stored in the DHT for a given user. Note
that cspace may be used on an operating system that supports teredo
natively, in which case it would appear to cspace that the host has a
public IPv6 address. This is good.

For hosts that don't have public IPv6 addresses running operating
systems that either don't support teredo or are configured to disable
teredo, cspace would have to include a userspace IPv6 stack and teredo
client. Using these, the cspace daemon could obtain an IPv6 address
independently of the rest of the system.

The biggest benefit is that cspace could take advantage of a capable
and well-tested system for NAT penetration without requiring
intermediate relays to carry all of the traffic or unnecessary
latency. A side benefit would be another step towards wider deployment
of IPv6. The reason I like teredo more than other IPv6 in IPv4 schemes
is that it doesn't require any configuration by the user and it
doesn't require the user to register with a tunnel broker.

Cheers,
Eric
_______________________________________________
cspace-users mailing list
[email protected]
http://tachyon.in/cgi-bin/mailman/listinfo/cspace-users

Reply via email to