Hi Robert,

On 09/13/2010 07:19 PM, Robert Hoffmann wrote:

> Hi there,
>
> a) I am not sure about what the practical benefit of hop-by-hop methods like
> CANCEL or ACK (for 3XX-6XX responses) is.
>
> b) Also, why is 100 Trying a hop-by-hop response? Would it hurt to have it
> being end-to-end?
>
> c) Why don't we have end-to-end messages only?

It seems to me that the logical purpose of hop-by-hop replies or 
request methods is to elicit acknowledgment or communicate receipt 
from all network elements in the signaling path, regardless of whether 
they are proxies or UAs, in situations in which that information is 
desirable to know from the point of view of semantics and the state 
machine.

So, for example, the standard envisions that it is useful to know that 
a proxy has itself received and processed a negative reply, even if 
the issuing UAC has gone away or something like that.

Hop-by-hop CANCEL is more or less a mandatory requirement of forking, 
as you mentioned, since in many cases the proxy is the conceptual 
target of the CANCEL request because the transaction is being 
presented to UAC opaquely as a single branch, while there are lots of 
branches on the other side to the management of which a CANCEL must be 
applied/translated.

The 100 Trying response exists mainly to dampen retransmissions by 
telling the UAC that the request has been received and is being worked 
on.  Sometimes, there are unpredictable sources of latency that arise 
at this stage of call setup if a proxy is involved, as, for example, 
due to:

- High database load / response time.
- High end-to-end network latency physically.
- Additional latency involved in serial forking through successively 
failing branches (common in wholesale PSTN routing scenarios).

Any one of these or other causes can cause Timer T1 to throw before an 
upstream 100 Trying from a UAS is received and forwarded, and cause 
undesirable retransmissions or premature failover.  So, 100 Trying is 
a way of telling the UAC, "All right, I got your request, now just be 
patient and wait, because there might be some latent operations 
involved before I can give you any other provisional or final replies."

-- 
Alex Balashov - Principal
Evariste Systems LLC
1170 Peachtree Street
12th Floor, Suite 1200
Atlanta, GA 30309
Tel: +1-678-954-0670
Fax: +1-404-961-1892
Web: http://www.evaristesys.com/
_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors

Reply via email to