On Mar 4, 2011, at 1:28 PM, Fred Baker wrote:

> On Mar 4, 2011, at 9:58 AM, Christian Huitema wrote:
> 
>> There are however two big limitations to solutions like STUN, TURN or ICE. 
>> First, it is entirely possible to engineer a middle-box that breaks them, by 
>> making the mappings hard to discover. Second, if the mappings are volatile 
>> and short lived, the end to end systems are forced into a pattern of 
>> frequent polling which drains the batteries of mobile devices. If the NPT 
>> spec addresses these limitations, we may well have  a deal. 
> 
> NPTv6 provides a stable mapping as long as the internal address remains 
> constant - it has an internal prefix and an external prefix, and does a 
> reproducible mapping between the two. There is a question with privacy 
> addressing - if the host interface changes its address, and especially if it 
> changes it frequently, the peer will need to keep track of that, and there is 
> a question of how DNS knows what the current address is. But the network 
> isn't changing the address, the host is.

I've always assumed that privacy addresses were meant to be ephemeral.   
Otherwise, there's no point to them.

Keith

_______________________________________________
nat66 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nat66

Reply via email to