From my reading of the admin guide (and someone please correct me if I'm wrong), the way a hunt through a trunk group is conducted is currently "shallow" in the sense that the hunt stops once it is determined the next port in the trunk group list is free to take another call, and that the number matches the DigitMap associated with that port. There is no guarantee that there is any service available on that port, however. If it is a SIP port, for example, an INVITE request might yet fail for any number of reasons.
On the other hand, a "deep" trunk group hunt would wait for verification that service _is_ available on a port before concluding the hunt. In the context of a SIP service, I think this would be at the stage of the call having progressed to "ringing" or "busy". At that stage, the service has proven itself to be up and running and available. On the other hand, if the connection to a SIP service provider, (whether a registered service or via a gateway or even URI dialling) does not proceed to the "ringing" or "busy" stage, then the connection attempt is deemed unsuccessful and the hunt should continue on to the next port in the list.
I appreciate this would not be trivial to implement, but my feeling is that this would raise the usefulness of the trunk group concept (which I think is a great one) up to the next level. Lets face it, no ITSP has 24/7 uptime, but the ability to automate some redundancy via trunk groups would be a great step towards creating "virtual" 24/7 availability. And I believe the feature would be truly unique for the consumer ATA market -- I think you'd have to go to something like an Asterisk box set-up to get this sort of sophistication otherwise.
Anyway, another feature request for the OBiHAI team's consideration. What do other forum members think?