Re: [sip-overload] draft-ietf-soc-overload-control-01 submitted

[Volker Hilt <[e-mail]>]

On 2/8/2011 11:37 AM, [e-mail] wrote:
>> -----Message d'origine-----
>> De : Vijay K. Gurbani [mailto:[e-mail]]
>> Envoyé : mardi 8 février 2011 14:56
>> À : CHATRAS Bruno RD-CORE-ISS
>> Cc : [e-mail]
>> Objet : Re: [sip-overload] draft-ietf-soc-overload-control-01 submitted
>>
>> On 02/08/2011 07:55 AM, [e-mail] wrote:
>>> [BC] I was trying to find a solution for the case where the proxy has
>>> been upgraded to support filtering of messages but not AS-A. I
>>> believe this will frequently happen in real networks (e.g. AS-A is an
>>> old system hosting basic telephony applications for user A while AS-B
>>> is a new system hosting a new overload control sensitive application,
>>> the core network proxies are upgraded).
>>
>> Bruno: I see.  The thing that bothers me about this approach --- and I
>> readily subscribe to your assertion that this may happen in real
>> networks --- is that (a) we are essentially performing overload control
>> for AS-B two hops away, and (b) we are attempting to codify B2BUA
>> behaviour in the context of overload.
>>
>> Even then, the proxy must have a filter so only those messages that
>> AS-A will send to AS-B will be quenched.  What is the nature of this
>> filter?  Is it prescribed anywhere?  Is loading this filter an
>> administrative task or will AS-A automatically send it to the proxy?
>> And if so, how?  These are all questions that arise in the context of
>> your scenario.  To what extent do we need to flesh these out if we go
>> this route?
>
> [BC] That's a bit tricky indeed. I think I understand you point as well. May be we should just add a warning that filtering must be performed by the adjacent nodes or is this stated somewhere already?
>
This is essentially going down the path of end-to-end overload control, 
which is discussed in the design considerations draft.

If AS-A is not updated to support overload control, it will drop the 
feedback from AS-B, which is exactly what it needs to do. If AS-A would 
blindly copy the feedback from AS-B to the proxy, it will throttle all 
traffic to AS-A even if that would not hit AS-B at all.

The proxy would have to be specifically updated to copy the oc 
parameter, which would be unwise to do. If an update is performed, it 
should be updated to support regular overload control.

Volker (as individual)