The claim (by Bell) is that a "third-party fibre line" way up in North Bend BC was the root cause of this problem..
I know providers all stick their own gear on the same towers, but I guess I didn't realize how common it was for the top-level providers (i.e. the ILECs) to lease cables (as opposed to running their own). I always assumed this was just the resellers (CLECs) who did it..
I wonder if that cable is fully third-party owned, or belongs to TELUS (the quad-play provider for western Canada) and Bell/Rogers lease capacity from them for their wireless services.
Bell and Telus got together to build their networks when they upgraded to 3G. Bell runs the towers in the east, and Telus in the west. (And they got Huawei to do it for them...)
https://www.theglobeandmail.com/technology/bell-teams-up-wit...
The CEOs of these competitors literally called each other up and had a friendly about how to save themselves a few billion while eliminating the risk of any price war between eachother.
They’re literally the same network appearing as 1.
So while you have the big3 « providers » and a bazillion of their sub-brands, there’s only 2 « national » networks.
And then Rogers leases some network space from Bell/Telus as the Rogers-Ext network for some rural areas, meaning everyone is using the same gear on the same tower.
Ideally you'd ensure geographically diverse routes for internet fiber, but I don't know how easy that is for legacy telephony networks.
https://www.google.com/maps/place/North+Bend,+BC,+Canada/@49...
OT, but I was curious and looked it up, and there are three North Bends in the Pacific Northwest (OR, WA, and BC). What a coincidence!
I guess emergency services must have a backup system or maybe there is a low bandwidth secondary that can't handle regular call volume and prioritizes emergency numbers.
The phone does not "dial" the emergency phone number. It tells the tower that it needs to initiate an emergency call. At that point, the infrastructure bypasses all authentication/authorization/billing checks and connects to what the infrastructure knows as the local emergency service. It's a standard voice connection, made over the same channels as a normal call, with one exception: if there is no available channel, it drops a normal call to make room.
In this case, some of the potential situations could be occurring: 1) There is still backbone (backhaul) availability, but perhaps the connection to the subscriber database has been severed. No normal calls can be authenticated, but emergency calls don't get authenticated and are thus not impacted. 2) Everything is working normally but at drastically reduced capacity. With the normal load of emergency calls, there is simply not enough capacity for normal calls. If this is the case, it is possible that not all emergency calls are being completed - the emergency operations center would not have the ability to know if this was the case.
The interesting implication of this is if I'm travelling in a country with another emergency number, I can "dial" the one from my home country (which is really just the sequence to initiate "emergency call") and be connected to local services.
I can see some references to 911 working on the UK phone system where the emergency number is 999 -- but this seems to be a separate mechanism (the phone company explicitly routes this number, as well as the EU emergency number 112).
There's a wireless version of this too although it doesn't quite work the same so I'm not sure how effective it really is.
[0] https://www.straight.com/life/1354586/pineapple-express-brin...
p.s. I’ve always liked that these winter rain storms coming into BC, Canada from the general direction of Hawaii are colloquially called “Pineapple Express”.