[asterisk-dev] Via/maddr: (was Re: [Code Review] Properly route responses according to the Via headers in the request)

[Kevin P. Fleming]

On 01/04/2011 05:26 AM, Klaus Darilion wrote:
>
>
> Am 04.01.2011 10:07, schrieb Olle E. Johansson:
>>
>> 4 jan 2011 kl. 09.34 skrev Klaus Darilion:
>>
>>>
>>>
>>> Am 03.01.2011 21:05, schrieb Simon Perreault:
>>>> On 01/03/11 12:30, Klaus Darilion wrote:
>>>>> Actually they can rather easy tell you where to send response - they
>>>>> just have to send from this certain IP address (in case of UDP).
>>>>>
>>>>> Thus I think maddr is a "bug".
>>>>
>>>> maddr is intended to contain a multicast address. In general you can't
>>>> just use this kind of address as source without breaking a ton of other
>>>> things (e.g. ICMP replies come to mind).
>>>
>>> Ok, this makes sense. But why should a response be sent to a
>>> multicast address at all?
>>>
>> If the request was sent to multicast, it's a natural result to send
>> the response there too,
>> to make sure that all recipients of the request got the response.
>>
>> Now, I haven't seen many implementations of SIP over Multicast. It is
>> used as a
>> test on every SIPit and makes all phones ring...
>
> So we could make a "SIP day" and ring all SIP phones in the world :-)

Actually, I think those tests at SIPit are done by sending INVITEs to 
broadcast addresses, not multicast addresses. A SIP UA would have to be 
explicitly bound to a multicast address to receive an INVITE sent to it.

-- 
Kevin P. Fleming
Digium, Inc. | Director of Software Technologies
445 Jan Davis Drive NW - Huntsville, AL 35806 - USA
skype: kpfleming | jabber: kfleming at digium.com
Check us out at www.digium.com & www.asterisk.org