CUBE HA with 2 ITSPs

[Phone_Guy]

I will be implementing a second ITSP provider on an existing HA pair of CUBEs. The new provider will be connecting on an available Gig interface. The existing ITSP is pointed at the WAN VIP from their equipment. For the second provider, do I just need to have them point to the WAN VIP as well, track that interface, and then just add that interface to the redundancy group, or is there some other HA related configuration I might be missing? 

[Roger Kallberg]

You need to connect the second provider to another interface than the current provider uses. If you want redundancy for the second provider you’ll need to configure a second redundancy group.

[Phone_Guy]

Could I just have the new carrier point to the existing WAN VIP?

[Roger Kallberg]

No. As the VIP is tied to the interfaces that the current service provider uses you cannot use the same for the second provider as it would be using another interface.

[Phone_Guy]

Right, that makes sense. I'm talking about not using a new interface on the CUBE at all. Just plugging the new ISP in to the WAN side switch and pointing it to the existing WAN VIP.

[Roger Kallberg]

What is the actual carrier technology on the side that you call WAN? Is it an internet connection where you terminate the current ITSP SIP service and potentially other type of series link SD-WAN and so on? If it is a internet connection are you planning on using the same connection for the new ITSP SIP service? I'm asking because in your original post you wrote this "The new provider will be connecting on an available Gig interface.", this part isn't very clear IMHO. Could you please elaborate on what you mean by that?

[Phone_Guy]

Below is a diagram and attached is the current config of ASR-1 with non-voice related elements omitted. Carrier 2 would be the new carrier. Although not shown, links are all going through a physical switch. Carrier 2 will be installing their own equipment and connecting to the switch.

 My assumptions:

Carrier 2 connects their equipment to the switch and I have them point at the WAN VIP.

- Allows me to keep the same redundancy group setup, but unsure of how to IP it. Can I have both carriers in separate subnets and still point at the WAN VIP?

Carrier 2 connects their equipment to the switch and then we connect new links from the switch to Gi0/0/2 on each ASR.

- Requires me to set up a new redundancy group or can I add to current group?

I feel like I have a handle on the call routing/dial-peer side of things, it's just this HA config with two providers i'm unsure of as I've never implemented it before.

[Elliot Dierksen · Accepted solution]

My off the cuff leaning would be to do either a fully separate interface or an 802.1Q sub-interface for the 2nd carrier. I would also put the 2nd carrier in its own VRF. Assuming the second carriers allows you a /29 or larger for the meet point network, you should be able to do the HA config as well. You likely still have a single point of failure at the carrier handoff, but you would have router redundancy.

[Phone_Guy]

After a lot of research this is exactly where we landed as well. We'll be using a separate interface in a different subnet with a VIP in that subnet. Add the interface to tracking, give it a distinct rii, and add it to the redundancy group. I tried this in my lab and it works as expected.

[Roger Kallberg]

Your VIP is from what I can tell on Gi0/0/1 and you say that you will connect the new SP to Gi0/0/2.

”Carrier 2 connects their equipment to the switch and then we connect new links from the switch to Gi0/0/2 on each ASR.”

How would you make that work with using the same VIP? Maybe I’m incredibly daft, but I just don’t understand what you’re trying to accomplish.