Does anyone know how to configure AAR?? 1

[dtrobert]

I've experimented trying to setup AAR to cause rerouting based upon insufficient bandwidth but haven't been successful.

I believe part of my problem is confusion about how they actually work. I realize the AAR group allows you to specify a prefix needed to dial the rerouted number. But what exactly triggers the call to route out the new gateway?

My first guess was to create multiple route patterns in different partitions. The AAR CSS would then be applied to the phone to cause it to match a new pattern in the event bandwidth was not available. Now I'm starting to think the AAR CSS is only used AFTER the new route is determined just to enforce a different class of service.

My other understanding is that gateways in a route group will automatically failover in the event of inavailability (lack of bandwidth or link down). In that case, why wouldn't one just assign some prefix at the gateway itself?

Any clarification would be great. I couldn't find much on cisco.com

Thanks

 

[ccmuser]

AAR will use your patterns that are already in place.. Thats why the aar setup asks you what you want to prefix.

If you use a 9.1xxxxxxxxxx (lonng distance)
and 9.xxxxxxx (local)

thats Pretty common so here are you basic steps

1.Set the aar prefix to 9 for local and 91 for between groups.

2.You need to change the service paramater to "allow" aar

3.You need to fill in your ext phone number mask on every phone(this is were it gets the full e164 add to make a PSTN call.

4. assign AAR calling search space

After all of these are done when CCM issues an ins bandwidth it will try to match a route pattern that is allowed in the AAR calling search space. and place the call via a matched route pattern.

Really pretty simple actually .. They just provide horrible docs on this!!!!!

Good luck

 

[dtrobert]

Thanks. Let me make sure I understand.

Support I have two sites (HQ1 and Branch1) with centralized call processing at the HQ1. I have a WAN link connecting HQ1 to Branch1 and am enforcing CAC using Locations.

Either the WAN link goes down or there is insufficient bandwidth to make a call from HQ1 to the DN at Branch1. (Say calling from 2852-HQ1 to 3660-Branch1). It's a long distance call over the PSTN to connect HQ1 to Branch1.

I already have route patterns for making local and long distance calls like you say 9.XXXXXXX, 9.1XXXXXXXXXX in partitions LocalPT, LongDistancePT.

I should now:
1. Create an AAR Group with the prefix of 91
2. Assign this AAR Group to the 2852-HQ1 line appearance.
3. Create a CSS for AAR (CSS-AAR1) containing partition LongDistancePT
3. Assign this AAR CSS to the 2852-HQ1 line appearance.

Now I think you're suggesting that when the WAN connection becomes unavailable, the phone (line specifically) will check it's AAR Group, add the prefix, and attempt to route the call once again. Of course, it is still missing the E.164 information to route over the PSTN to the other Branch1.

To provide that information, it will utilize the External Phone Number Mask on the phone I'm trying to reach. So if line 3660-Branch1 has mask 415543XXXX or even 4155433660, then the original call will dial 91 + 4155433660.

1. Am I correct?
2. Why do trunk's themselves have AAR Group/CSS settings?

Thanks a lot.

 

[ccmuser]

ya pretty close.. Depending on your use of link is down it may or may not work. You may want to look into using route groups and route lists to handle GW in this scenerio but for CCM registered IP phones what you are saying makes sense.

You assign lines to AAR groups and the WHOLE phone to AAR CSS. In theory can give you AAR CSS permision to call everything if you only have one CSS on the phone.. The call has already been approved by primary CSS. so you are really just doing GW selection here.

Your assumptions about call flow are right on!!!


Trunks and GW have AAR groups for Out of BW situations just like phones, the main application for this is AA funtions on an inbound gateway where the destination is unreachable. the GW has to transfer the call to Voicemail here. So and AAr situation is needed.(this is because unity is seen to CCM as a skinny device).

CSS search space really applies along the same lines, But only if you are using mulitple partions for your internal phones and GW.

 

[dtrobert]

Thanks again.

So are you saying the AAR CSS will be checked first, before the phone CSS for the route pattern selection (kinda like the DN CSS)?

The situation I envisioned was where a company wanted to restrict users to local calls only plus any branch office. In the event there was no BW available, they would grant them more privledge to place the long distance call to the remote office but only during AAR.

You mentioned using Route Groups for failover rerouting which brings up another confusion point.

If I have a gateway with say a WAN and PSTN interface, and the WAN interface goes down, where do I tell CCM the prefix to use to reroute the call out the PSTN? My route group would essentially consist of WAN port, followed by PSTN port (MGCP). It really doesn't matter that I have a single gateway, the same question would apply if I had separate WAN router and PSTN gateway.

Thanks

 

[ccmuser]

Ya.. I think you are kinda confused on this one.. The Route pattern is going to point to a Route list.. A routelist points to routegroups..Which point to your individual ordered Gateways.. Either Ip or PSTN...

The AAR CSS is only used AFTER the "OUT of BW" message is given. So it is essentially checked after.(but the order really would not effect the result)This So you can only let certain users use the PSTN when wan connecions go down..

 

[dtrobert]

Okay, ya I know about route patterns->lists->groups->gateways. I guess my new questions isn't necessarily related to AAR.

Let's assume I am using toll bypass to route calls in my HQ local area over the WAN between my branch and HQ. At my Branch, any call that matches say 91415XXXXXXX will be transformed using a mask XXXXXXX and use the HQ PSTN gateway because it is the first gateway in the route group of the route list for that pattern.

I would also want the ability to route those 91415XXXXXXX calls out my branch gateway as toll calls in the event the WAN went down. So I figure I would add that gateway to the same route group but in the 2nd position.

In the case where I am routing out my 2nd gateway, I don't want to strip the 1415 any more yet my route pattern transform is set to do this. I would think the gateway itself should have a transform but don't see it.

Thanks

 

[ccmuser]

You are really better of creating a separate route group for gateways of the same Priority(ie if you don't care too much wich if these you use.(ie if you have 4 pri's at your main site.) Put all of those in one RG.

If you have individual GW at your remote sites put each of those in it's own RG. Then build Route lists for Each individual Route-pattern you want to route call for.

What this means is that you need separate router patterns and partitions for every class of service you want to provide in every campus.

Ie. 2 campusus with different call routing needs means you need Seperate CSS's partitions and Routelist's for every one.. so if you want to offer 911_local and Long_distanse CSS spaces, You need 4 partitions, 4CSS and 4 routepatterns and maybe 4 routelists. You don't really need any more Route groups though.

 

[dtrobert]

Okay so for the scenario I described, you're saying to have two route lists->route groups, one for the HQ gateways, 1 for the Branch1 gateway.


Now, for Branch1 phones, I want to be able to have my HQ gateway tried first, before going to the local Branch1 gateway if the call is local to my HQ.

I'm not following how you mean to create two separate Route Patterns to allow for this type of failover.

I can create 9.1415XXXXXXX and have my Branch1 phones use it to route calls over my WAN to the HQ gateway. However, if my WAN fails, how will they know to try another Route Pattern that points to the Route list containing my Branch1 gateway?

 

[ccmuser]

Ok you will have a RP 9.1xxxxxxxxx in partition Locatioan_A_LD. This RP points to a Route list Location_A_LD_routelist that points to RG that in it is an "ordered list" pointing to the "gateways" you want to call.

You will also have a RP 9.1xxxxxxxxxx in partition location_b_ld. This RP will point to a different route list that uses a diff. ordered list of Gateways.

I hope this clear things up a little.
I used the terms "ordered list" and gateways a little loosely but I think you should get the point.

 

[ccmuser]

Oh BTW this means you need two separate CSS's for each location

 

[dtrobert]

Okay, so my route groups would have the HQ gateway both in the primary spot, the Branch1 would really differ in that it would have an additional gateway (the local one) in the last priority spot, right.

Then I just point my Branch phones to it's Route List/Group via different CSS.

Sounds good. Thanks

 

[ccmuser]

That works..... Assuming you want all Ld calls from both locations to go out HQ gateway first.

But I would do it like this.. create a route Group for HQ resources. Create a seperate RG for Branch one resourses.

Then use your Route lists to controll call Flow. Ie a route list for PRI selection for your remote site pointing to HQ routegroup first then to remote routegroup.

Just easier to manage when you have large blocks of the same type of resourse at multiple locations. If you only have one pri at each location then I think your method will work... But is never a bad idea to get into good habits that may make your life simpler later on as your route plan grows.. it becomes much easier to change and alter call flow this way.