IPO Server Edition SIP Trunk Redundancy

[AvantiWade]

Hello All,

I am trying to understand how to implement a truly redundant IPO Server Edition. I have setup server edition and a tinkered around with SIP Trunks but I am not grasping how to create a fully redundant solution. I understand and have demonstrated 960X phone resiliency by powering off the system while a call was in progress as well as a fail over to a Secondary Server. Where I am lost is on the PSTN side of the system.

In a redundant topology would I need to register the SIP Trunk to both the Primary and Secondary and work with the provider to ensure this works correctly? Is it even possible to have separate systems registered to the same SIP trunk?

The avaya documentation only discusses phone resiliency, not the trunking portion. I haven't found any related posts on tek-tips or the net.

I'd appreciate any guidance the community can provide.

 

[pkdev]

Hi,

To have a fully redundant IPO set up you will have to register the provider SIP trunk to both Primary and Secondary.
This means you will need to have twice SIP Trunk License.
You also need to know if your SIP provider support this set up (it should).
A good ARS configuration will do the backup job.
To backup incoming calls you will have to create all the incoming route on both primary and secondary

Depending on how the provider is connected to your IPO set up you will even be able to have call preservation during primary server failure

 

[AvantiWade]

Thanks for the reply pkdev. Is there any specific documentation on the Partner Portal that you could link me to read up on? I was under the impression that all licenses were shared across the Primary and Secondary, but in this scenario we need to apply SIP Trunk licenses to the Secondary?

 

[hibroth]

Officially the SIP trunk resiliency is not supported on the SE, so the only solution is the one suggested by pkdev

 

[AvantiWade]

I don't mean to be brash but I am failing to see the benefit of a Secondary Server if there is no inherent resiliency for the PSTN side of the system. In a scenario where a customer has a server failover event the phones will re-register after a few minutes to the secondary system but they won't have any external telephony, only internal between extensions? This seems like half baked resiliency.

Would a better topology be having a IP500v2 as the termination point for SIP trunks and point the SE's towards the IP500v2 with h323 trunks for PSTN access? It still results in a single point of failure but at least during a server failover event they would have PSTN access.

 

[derfloh]

But if the IP500 breaks they have no more PSTN access.

In my opinion if your provider supports a double registration of that SIP trunk you will be fine. You can also have an ISDN trunk shared over two PRI lines (60 channels) and be able to connect these two PRIs with two IP500s. In both scenarios the proper function depends on how good you program inbound and outbound dialing. Don't forget to think about special connections like analogue devices connected to IP500, fax server and do on.

With SIP you have to buy as much trunk channel licenses ad you want to be able to use simultaneously. Independent if you use it with primary or secondary server or if you share the channels over both servers. They have to be licensed against primary server system ID.

 

[nothingworks101]

our sip provider in UK said we cant have 2 systems registered at same time. that I would need to have secondary programmed but off, then activate manually when the primary server goes down.

completely not useful.

 

[derfloh]

Then it is Not the IPO's fault.

 

[vztech]

Avanti

Primary - SBC -SIP (lets say 30 channels) to carrier A
Secondary SBC -SIP (lets say 30 channels) to carrier B

The only issue I see if you used carrier A&B would be the incoming DID's. I'm not sure if carrier A fails you could swing your DID's to the second carrier B so you might need to go with the same carrier to both systems.

ARS that utilizes both A & B SIP trunks

Done !

It's only dialtone!!!

 

[AvantiWade]

Thanks for the info and discussion, I'm going to spin up another demo system and try to lab this out with a test trunk.

Without getting to far off topic, how are all of you ensuring Communicator, One-X Mobile, and other soft clients reconnect to the secondary server after a fail over? DNS Trickery or just telling the users to change their settings from "primary.company.com" to "backup.company.com"?

 

[derfloh]

Perhaps a SBC can do that. If you want to be sure it can be possible to build a HA cluster out of two SBCs. But to be honest I have no experience with that scenario.