I’ve seen if there’s too much buffering or some sort of other issues it drop
out and not work. I’ve actually replaced customer CPE to address this.
You want to make sure that you are passing ESP through, and there’s even issues
with some of the carriers not liking a phone using wifi calling if it has a
public IPv4 address (!)
I don’t recommend this, but will point out that the carriers have defined SSIDs
you may be able to shunt the devices over to, and if you do the open roaming
stuff it may also help depending on your environment.
ATT_US.bundle/profile.mobileconfig- <string>attwifi</string>
Verizon_LTE_US.bundle/profile.mobileconfig- <string>VerizonWiFiAccess</string>
You also see this in their config for their SSIDs:
<key>DisableAssociationMACRandomization</key>
<true/>
If you have BYOD but under enterprise management it may make sense to shunt the
phones to a discrete SSID which has fewer restrictions.
- Jared
> On Aug 2, 2024, at 11:10 AM, <chuckchu...@gmail.com> <chuckchu...@gmail.com>
> wrote:
>
> Hey all,
> Question if anyone knows about cell phone wi-fi calling in
> US. Googling isn’t finding what I’m looking for. We have a corporate site
> in US where users have BYOD capability, and use their phones with wi-fi
> calling enabled. Site uses a single NAT address (IPv4) for BYOD access.
> Recently the site reported wi-fi calling had stopped working for all user
> phones, Apple and Android, all about the same time. The guest network did
> have some bandwidth limitation applied and they had overuse. That was since
> resolved, we upped the bandwidth. But the phones all still avoided wi-fi
> calling. It’s a manufacturing site with known cell signal issues, so most
> users had no signal via carrier. I did not get a packet capture yet to see
> what could be going on, we’re 99% sure we’re not blocking traffic. I’m
> wondering if the phones have an algorithm to test wi-fi signal, and perhaps
> the carriers will blacklist public IPs with known wi-fi calling issues to
> avoid cases where an emergency call can’t be made because of intermittent bad
> performance? It seems odd that even when no bandwidth issues exist, it’s not
> attempted.
> Thoughts?
> Thanks,
> Chuck Church
