On May 9, 2008, at 3:24 AM, Juha Heinanen wrote:

> Dean Willis writes:
>
>> But how does a UA know in advance that a proxy is going to send it?
>
> it doesn't need to know in advance.
>
>> And how does the UA know that it is coming from a proxy in response  
>> to
>> its own CRLF stream, and is not an artifact of an echoing NAT or  
>> other
>> weirdness?
>
> so you assume that nat or some other network box sends back on its  
> own a
> single CRLF when UA sends double CRLF to its proxy and in that  
> response
> all tcp layer stuff would match?  in my opinion trying to cope with  
> that
> kind of "weirdness" goes too far.  how do you know that your nat box
> didn't generate on its own a 200 ok when you sent invite?


Barring very clever timing and counting, it's hard to tell whether a  
CRLF stream  coming back to a a UA is really single CRLFs in response  
to double CRLFs or is CRLFs being echoed from a broken firewall.

And yes, I have seen firewall boxes get confused and echo stuff back.  
I've also seen TCP stacks get in strange loopback states where the  
echo is "local" rather than coming from a NAT box. I've never yet seen  
one get confused enough to send a 200 OK that matches an INVITE, but I  
suppose somebody could make one that does someday. Never underestimate  
the power of stupid people in large numbers.

--
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