I have a CCM v9 Cluster in a campus environment. We are migrating from PRI's to SIP Trunks. We will have 4 SIPT's terminating in 4 different locations. Our SIPT's will be configured using session pooling. Essentially, all 4 SIPT's are aggregated. We can make/receive calls from any location out/in any SIPT.
CCM is configured with a dial plan that utilizes the local-route-group concept. I have 4 SIP gateways (CUBE's) added to my CCM. I have 4 local route groups, one for each site where a SIPT is terminating. Each route group will have an ordered list of the CUBE's I have added. I have only 1 route pattern for all external calls.
This is where I am having a hard time conceptualizing this. When a phone dials an external number, based on the phones device pool, it will use the ordered list of gateways in the route group. My external call hits the single route pattern I have and routes it out the first CUBE it has been assigned via the phones device pool/LRG. However, I am trying to implement CAC/AAR to my SIPT's.
After reading about AAR, CAC, etc., the call only uses an AAR when location/bandwidth is full. So, in the event that the first CUBE in the route group is down, the second CUBE in the route group is used for the outbound call, the third, and so on. That is correct, correct?
So, to understand AAR, is the location bandwidth calculated based on the calling device and the first CUBE in the route group?
Assuming that it is, how can I reroute to a backup path in my scenario? Would I have to setup a second route pattern (only accessed when AAR is invoked), and then use a route list for that 2nd route pattern that only sends the call to one particular CUBE (no route list)? What if that bandwidth is saturated, can I route to a tertiary CUBE? What if I used another route list in my second route pattern, then how is the location bandwidth consumption being calculated? Assuming the first CUBE in the route list on the 2nd RP will be different than the first CUBE from the route group assigned to the DP, the bandwidth is calculated with that second CUBE?
This is a hard example to articulate via words.
Ultimately, I am not going to use AAR to reroute a WAN called out a PRI. I am trying to reroute an external call (WAN) out a different SIP Trunk when the numbers of calls to that CUBE/location are close to exceeding the total bandwidth/session capacities.
