Asterisk + POTS + VoIP

Hey guys I got a few questions about asterisk. I know linux pretty well so I am not worried about that. I needed to verify some stuff. A yes or no answer will suffice.

  1. I have one POTS line. Can that POTS line be hooked up into my Asterisk box and calls get forwarded from there to the phones? (The phones will have ATA’s attached to them GT-286)

  2. I would like outgoing calls to be routed though my VoIP service provider instead of the POTS line. So, incomming is POTS outgoing will be VoIP. Is this possible?

  3. If my internet connection fails can asterisk switch all the calls to POTS? for outgoing calls?

  4. Can I have two VoIP service providers? Ex: ABCvoip & XYZvoip. So lets say all my VoIP calls goes through company ABCvoip, all of a sudden they are gone and out of business can asterisk automatically switch over to XYZvoip?

Thanks for your time, Any comments or suggestions is truly appreciated.


  1. Yes
  2. Yes
  3. Yes
  4. Yes


Thank you very much for your time. I really appreciate it.


Regarding 3 and 4. Asterisk can automatically retry a failed call by a different route. However it is difficult to reliably distinguish between calls that fail becauses of network problems and calls that fail because of end user problems, so there is a significant risk that you wil end up making multiple calls to the end user, when all your networks are working.


Is there an order than can be set? For example. If the system cannot contact

VoIP company #1 then try
VoIP company #2 if you still can’t dial out
Go over POTS ?


It works at a lower level than that. You really need to read up on Asterisk dialplans, but basically, you do a wild card decode of the dialled number, qualified by a context that depends on where the number came from. You then perform sequence of operations until the call succeeds or you run out of options.

One of those operations is Dial. In Dial you specify, speaking loosely, either a local phone or an outgoing trunk, and the number to send to that trunk. The number can be derived by a string manipulation of the one that matched the wild card pattern.

If you want to do fallback, you alternate Dial with logic to decide whether the failure was in the network or with the subscriber. That intervening code can be empty, but, if the callee doesn’t answer, they will end up with a missed call for each of the primary route and the various fallback routes.

As these rules are interpreted sequentially, you can try different carriers in any order that you like.

(The typical use is to implement department hunt groups, or falback to voicemail, in which case one probably wouldn’t check the detailed failure cause between each step.)

More generally, a lot of things are done at a lower level on Asterisk than a typical PABX.


I am currently reading a book called Asterisk: The future of telephony (second edition) They talk about running asterisk 1.4 I think this book was written a few years back. Even though it might be alittle outdated but I think it will give me something to start with.

Right now I am thinking I should have a asterisk server in each one of my locations and use solid state hard drives.

I have one more question for you (or anyone). Do you think it’s a good idea to have a asterisk box in each location or one central one?

I mean I need phone redudancy with POTS so I am think the only way to achive that right now is to have asterisk in each location.

Again, I don’t need a detailed explaintion just a yes or no with a short summary is good. I am just putting together a proposal for my company and within my budget I have asterisk support =) in the cost factor. But I would like to get your thoughts. Or anyone willing to give me their 2 cents.

Thanks a million

I’m on the development side, rather than the operations side, so it is maybe better to let someone on the operations side answer. However, the tradeoff is probably been simlicity and trunk traffic volumes.