The constraint is a single PSTN path per-logical location in Control Hub. So yes, this is possible.
If the actual PSTN circuits will remain on-prem using the Local Gateway approach you have at least two options:
- Configure separate locations and trunks in WCH, one per-subsidiary, and then use dial-peer groups on the gateway to dictate the outgoing dial-peer based on the incoming dial-peer of a call. This will likely require unique interfaces though (loopback is a valid option) so there are separate TCP sockets per-trunk. Maybe also unique certs if you don’t use the registration approach; I haven’t tried that.
- Use a single trunk in control hub with either a single location or per-subsidiary location and sort it out on CUBE instead. The incoming dial-peer would probably need to match based on answer address (ANI) instead of the newer and more definitive SIP URI method though. From there you could use dial-peer groups to dictate the outgoing DP.
Note that Survivability Gateways are limited to a single location, so consider that constraint if you need fallback. I’m not sure you could maintain PSTN separation during SGW fallback though; I haven’t needed to think through that.
If the intent is to migrate to a Cloud Connect PSTN provider to eliminate the on-prem dependency, each subsidiary intends to use the same provider, AND have separate accounts/contracts you will need to check with the provider if they can support that. Same provider account or different providers per-location would be an easy yes.