[asterisk-dev] What's the best source for architectural understanding?

[Eric "ManxPower" Wieling]

For SIP, a call comes into Asterisk and is handled by chan_sip.  The 
call is authenticated against sip.conf.  The context= line for that 
device is the context the call falls into in extensions.conf.  If the 
call came in with the destination number provided then Asterisk will 
send the call to the matching exten => line in extensions.conf.  If no 
destination number is provided (and ONLY if there is no destination 
number) then the call is sent to exten => s.

It pretty much works the same for all the protocols.

Jay R. Ashworth wrote:
> On Wed, Oct 04, 2006 at 12:32:44PM -0500, Moises Silva wrote:
>>> There seems to be quite a bit of Magic on issues like call control
>>> hidden inside the 'Applications' one can call from a dial plan
> 
>> No magic here, check this out:
>>
>> http://www.lobstertech.com/doc/ast-12-func/
> 
> I did.  That's back down at 50 feet -- if not below ground -- telling
> how to *breat* functions.
> 
>>> (and indeed, the way by which calls *get to* a dialplan in the first
>>> place; is there a good overview of these topics which, for some
>>> reason, I've been too stupid to locate? :-)
> 
>> Basic SIP setup. The general concept apply to any technology (IAX2, Zap 
>> etc).
>> http://lists.digium.com/pipermail/asterisk-dev/2006-September/023116.html
> 
> And that's at 100 feet.
> 
> We're still nap of earth there, and I'm too new to read between the
> lines yet.
> 
> What I'm looking for is
> 
>  "A monitor process is active within the Asterisk core for each defined
>  channel; when a channel receives a connection (from an incoming call
>  or from an internal extension being picked up to place a call), the
>  monitor process begins executing the dialplan for the appropriate
>  context, which is X for "trunk connections" (things which present a
>  ringing signal or equivalent) and Y for "station connections" (things
>  which present an off-hook condition).
> 
>  "When a call is extended through the switch to a destination port, the
>  call monitor of the originating port controls the call; the destination
>  port's call monitor sleeps."
> 
> ... or whatever the actual facts are.  
> 
> What I'm saying is, I can't visualise the core architecture of call
> control in an Asterisk switch... which as someone who's planning to
> start playing pretty hirsute games with one, makes me nervous.
> 
> Is *that* layer of documentation around anywhere?
> 
> Cheers,
> -- jra