Dean Willis wrote:
> 
> On May 8, 2008, at 2:27 AM, Jiri Kuthan wrote:
> 
>> Juha Heinanen wrote:
>>> Christer Holmberg writes:
>>>
>>>> I agree with Hadriel. The idea is to let the client do it, so that
>>>> the proxy doesn't need to handle it, e.g. by providing short
>>>> registration refresh timers etc in order to make sure the NAT
>>>> bindings stay open.
>>>
>>> christer,
>>>
>>> could you then, please, restrict the scope of your draft to udp
>>> connected UAs, because if they are tcp connected, clients can keep
>>> their nat binding open by simply sending double CRLFs.  if they get back
>>> single CRLF, then fine, but if they don't, it is fine too.
>>
>> True, I didn't realize that, with TCP we don't even need the reverse 
>> CRLF.
> 
> Well, without the reverse CRLF, you may not notice for 20 minutes that 
> your connection is down. The TCP stack will take a while to timeout from 
> SYN.

Well, that's compelling enough indeed to have CRLF back and forth.

-jiri


> 
> -- 
> Dean
> 
_______________________________________________
Sip mailing list  https://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [EMAIL PROTECTED] for questions on current sip
Use [EMAIL PROTECTED] for new developments on the application of sip

Reply via email to