On 2010-04-21 at 10:39 -0400, Edward Ned Harvey wrote:
> This is particularly interesting, because IPv6 doesn't bring much to the
> table over IPv4, except in the area of enabling p2p communications.  

Publicly routed IP addresses, as many as you want.  Wonderful when you
hit things like Kerberos with matching forward/reverse address
requirements.

Also avoids the need for multiple layers of NAT, which is what various
areas of Asia are looking at needing.  Providing carrier-grade NAT is
expensive and will have per-source-IP restrictions on parallel
connections.  I saw a presentation a couple of years ago on how bad
Google Maps looks, in terms of tiles which don't render, as you adjust
the permitted number of parallel connections.  I believe this was at the
2008 Google-hosted IPv6 implementers forum.

Thus the incentive for people in countries which don't have the surfeit
of addresses that the USA has and the incentive for any providing
service internationally.

> I can think of at least one really significant way that NAT-PMP or a similar
> protocol would be more desirable than IPv6.  Namely, in order for the
> inbound port to be accepted via NAT-PMP, the internal client has to
> specifically request it, and it's only valid as long as the client maintains
> the lease with the firewall.

Right, you can have firewalls with IPv6 too, with the same default-deny
policy.  At this point, you're into the "middlebox" terminology.

This is actually independent of NAT.  After all, a few years back, when
NAT boxes started being produced, it was common for IPv4 stacks to
support source routing, which neatly bypasses NAT's supposed security.
The inbound security comes from the co-hosted firewall and turning off
source routing support, not from NAT itself.  NAT does help inhibit
discovery, but that just deters casual attacks.

It would be nice to see a simple protocol, along the lines of NAT-PMP,
which lets client devices punch small holes in firewalls.  Of course,
malware will just use the same client APIs, so at some point you should
ask "why do we bother?" but in terms of protecting misc devices in
shared subnets, it may help a bit.

> Unfortunately, it appears not many firewalls today support either NAT-PMP or
> IGD.

Consumer-grade devices commonly support IGD, a few also support NAT-PMP.

Corporates tend to not like the idea of random desktops punching holes
in the firewalls.

-Phil
_______________________________________________
Discuss mailing list
Discuss@lopsa.org
http://lopsa.org/cgi-bin/mailman/listinfo/discuss
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to