Tricky Design Question? Is it Possible?

[brian-henry]

Hello to all,

I have been tasked to make this work and do not know if it is possible and if so will my idea work or flop! ļ

Sites

Site A ¡V Main HQ ¡V 8PRI Ld/Local

Site B ¡V Remote Site ¡V 1 PRI/4FXO Local

Site C ¡V Remote Site ¡V 1 PRI/4FXO Local

Connection b/w HQ/Remote Sites ¡V GigE(Primary), Backup T1

Problem:

If Site B or C wants to make a LD call then it must go out the PRI¡¦s at HQ

Connection(S) UP ¡V This is easy its just RL saying go out the HQ PRI(s)

If Connections DOWN ¡V Tricky

Calls must go out the Local PRI dial into the HQ PRI(s) and then dial back out to complete the LD calls? I thought that maybe if on the route list for the Local PRI calling it LDFO-LOCAL-HQ and having it when it dials out the Local GW have its calling number set to 1 specific number so when it calls HQ I can have an Inbound Translation Pattern that says I see that and my CSS will only allow me to use the LD RP¡¦s at HQ?

If the Connection is down b/w the sites and the Local PRI is down at the Remote site then LD calls will go out the FXO ports dial into HQ PRI¡¦s and then back out?

Is all this possible or am I just spinning my wheels trying to make this work?

Thanks for any and all responses to my current question.

Brian

[RICHARD MESSINGER]

Two possibilities:

1) If you have a centralized Call Manager where the cluster is only at the HQ, then you are looking at configuring SRST on the routers at the remote locations to take over the call handling when the connection back to the HQ goes down.

2) If you have a distributed Call Manager cluster, where there are cubscribers at each of the remote locations, then you will need to have in the ordered route list to go to the route group for going out the remote gateway to do calling number manipulation.

[brian-henry]

Thanks for the reply,

Each site will have its own CCM for its devices and I worked with SRST but this is more of how to route calls then when in failover mode of lossing contact from a central cluster design.

The thing I guess I was tryin go get at was has anyone or could it be done of:

Hairpinning calls that come from one source and have it go back out the other source but maintain the call leg and not drop the call.

So the route would look like this.

If the GigE/T1 is down and your local PRI is up along with local FXO ports

Site B LocalGW-LDCALL->SB-LDRP->SBRL-LDSB-HQ->HQPRI(IN)->TLR->HQRP-LD->HQRL-LD(OUT)

The RG on Site be would have its calling number be set to (ONE VALUE) that could be put into a translation rule and then have it redirected back out the LDPRI at HQ. The same could then be done for the FXO, but it would be the (THIRD) CHOICE in the RL selection.

Thanks again!

[paul.harrison]

Brian,

Getting the router in SRST mode to make all calls to the HQ PRI is not a problem, just configure the dial peer to translate any number to a single number at HQ.

Getting onward calls to LD destinations is a lot more tricky. How are you going to dial the LD destination? You are probably only going to be able to achieve this using an IVR script which answers a call from the remote site, checks the cli then requests the onward number once it knows the caller is genuine.

You could use this method for all calls by using multiple CTI route points, so outbound calls by dialing 9 hit a version of the script which does not check the CLI, but calls into a specific DID number do check the CLI, then just get the users used to hearing a prompt after they dial 9.

This is probably the most elegant way of meeting your requirements, but I reckon the cost of implementing the IVR would be so great that any cost savings you might get by dialing all LD calls via HQ will be swallowed up and then some.

Paul