How to load balance calls when Gateway have 2 PRI

[actualabhishek]

Hello,

I have Gateway VG01 with 2 PRI from ISP 1 and ISP2. Both these PRI are currently member of Trunk-Group PSTN. This gateway is defined as H323 on CUCM side, so there is no control on ports.

With current setup only 1 PRI trunk from ISP1 is in use (for outbound calls), since its a E1 trunk with 30 channels for call and not all channels are ever used so 2nd PRI never get hit.

Can someone suggest if we can change hunting withing trunk-group to load balance calls. I want hunting similar to circular algorithm in CUCM.

Thanks

Abhi 

[Ratheesh Kumar]

Hi there

You can create a  RG with both H323 GWs, then use the distribution algorithm as Circular. Incoming you can talk to your Telco for load balancing between different PRIs.

Hope this helps!

Cheers
Rath!

***Please rate helpful posts***

[R0g22]

You said like CUCM, how can CUCM route to specific PRI's on an H.323 gateway ? It cannot.
From a GW perspective, I don't believe there is a way to achieve pure load balancing. Trunk group is an option but it's more or less like a PRI ether channel with hunting enabled which serves the purpose for failover purposes.
An option could be to create multiple dial-peers per destination and have "no huntstop" enabled. Again this IMO, isn't a pure load balancing mechanism.

[actualabhishek]

I referenced CUCM hunting just for an example. I believe load balancing can be achieved by using trunk-group hunt scheme round-robin as per below Cisco documentation

https://www.cisco.com/en/US/docs/ios/12_3/vvf_r/vrg_h1_ps1839_TSD_Products_Command_Reference_Chapter.html#wp1015378

[R0g22]

Nope. This is what it says -

For example, suppose a trunk group has three trunk group members: A, B, and C. Trunk group member A has the highest preference, B has the next highest, and C has the lowest. The software starts the search with A:

•If A has an idle channel, that channel is used, and the next request for an idle channel starts with B.

•If A does not have an idle channel, the search moves to B:

•If B has an idle channel, that channel is used, and the next request for an idle channel starts with C.

•If B does not have an idle channel, the search moves to C:

•If C has an idle channel, that channel is used, and the next request for an idle channel starts with A.

•If C does not have an idle channel, the search returns to A.

If an idle channel is found in a circuit (let's say A, it will use that and only when an idle channel is not available in A circuit, the search moves to B). This is NOT load balancing. You could potentially achieve this with a "no huntstop" config that I highlighted as well if that is all that you are after. This is a failover mechanism. You can trying to load balance.