Re: [RRG] Six/One Router Design Clarifications

[Brian Carpenter <brian@cs.auckland.ac.nz>]

[sending again from an alternative address - gmail is
being rejected by the RRG list again, and the whitelister
refuses to whitelist gmail this time]

Hi Christian,

On 2008-07-24 05:58, Christian Vogt wrote:
> Brian,
> 
> sorry for responding with delay.
> 
> I do agree with the disadvantages of translation that you are naming.
> But let me re-emphasize that my arguing in favor of translation is on
> the basis of this being a deployment tool, not a permanent mechanism.
> 
> I believe we agree that, of the three issues with NATs,...
> 
> (1) non-unique addresses make it hard to contact hosts behind a NAT
> 
> (2) state inside the network prevents re-routing of traffic
> 
> (3) non-end-to-end addresses require NAT traversal by applications
> 
> ...the backwards compatibility mode of Six/One Router avoids the first
> two because it is stateless and uses only globally unique addresses.
> 
> What remains is issue (3).  I believe we agree that no future routing
> architecture should have this issue on a permanent basis.  Where we
> disagree is whether the issue is acceptable for backwards compatibility
> and deployment.
> 
> Lixia has named it:  Whether we deem issue (3) acceptable or not, the
> issue will be present in IPv6 anyway because hosts will have to
> traverse IPv4/IPv6 NAT-PTs.  The functionality that applications
> *require* because of NAT-PTs is exactly the functionality that Six/One
> Router's backwards compatibility mode will re-use.

Yes, even with the replacements for NAT-PT that are being
discussed over in V6OPS and BEHAVE, the underlying NAT issues
of ALGs, helpers, and port mapping will remain. My concern is that
once we admit those code paths to a pure IPv6 world, we will never get
rid of them and the path to NAT CPE will be open.

> Furthermore, the time period during which NAT-PTs will be needed will
> likely outlast the time period during which Six/One Router's backwards
> compatibility mode will be used, due to the yet-small deployment of
> IPv6 compared to that of IPv4.  (This is a personal opinion, though.)

I agree.

> 
> Brian, I fully understand and agree with the concerns you have with
> translation.  Where our opinions differ is the extent to which
> translation is acceptable as a deployment aid.

Yes
    Brian




--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg