Re: [Int-area] practical issues with using v4-mapped addresses for nat64

[Iljitsch van Beijnum <iljitsch at muada.com>]

On 26 jul 2008, at 6:10, Pekka Savola wrote:

> > No, if there's no translator the packets won't end up at the  
> > destination because the mapped prefix isn't routed on the v6  
> > internet.

> If it's the destination address -- yes, but you might be surprised  
> how widely default routes are being used.  Enterprise-ISP egress use  
> them a lot.

Right. But nobody configures mapped (or compatible) addresses on their  
host interfaces, and nobody puts them in their routers. So even if  
packets travel a long way by virtue of default routes, they're not  
going to be delivered to the right destination successfully unless  
there's a translator in the path that is set up to translate packets  
with these addresses.

> The result is probably that packets end up looping between two  
> routers at some ISP's backbone until the hop count rearches zero.

No, that generally happens when a customer has a prefix, such as a / 
22, but they only use part of it, like a /23. If they then don't  
filter out packets to the unused space (the other /23) the packets  
will flow back to the ISP through the default route and the packets  
loop.

In this particular case ISPs won't send these packets to customers.  
They'll follow default routes until they reach a router that doesn't  
have a default or until the packets are filtered.

> If it's the source address (it wasn't obvious whether this would be  
> used, but apparently at least in translator->v6host direction),  
> there is no check from this POV.

Right, from a translator to a host the packets could be delivered if  
they're not (ingress) filtered.
_______________________________________________
Int-area mailing list
Int-area at ietf.org
https://www.ietf.org/mailman/listinfo/int-area