[asterisk-dev] The New CDR system

[Rod Dorman]

On Friday, March 30, 2007, 19:29:38, Russell Bryant wrote:
>   ...
> Well, the way I am thinking about this, absolutely no "conversion" would
> be taking place within chan_iax2.  The raw event would just be queued up
> into the core and then chan_iax2 would move on.  The amount of 
> processing would be roughly equivalent to that of queuing up a single 
> audio frame.  Meanwhile, in another thread, the event would be 
> dispatched to all of the local subscribers to that event type.  This is 
> where the actual logging to the database, writing to a file, or 
> whatever, would take place.

Is IAX always going to be over TCP?

If  not, presumably there will have to be some reverse traffic either to
send  an  ACK  after  some  window size number of events or to request a
resend of a missing event sequence number.

-- 
rodd at polylogics.com     "The avalanche has already started, it is too
Rod Dorman              late for the pebbles to vote." - Ambassador Kosh