5 ms·
One important element would be to rotate the hardware identifiers frequently, and, if the network is collecting payment from its users, totally decouple billing
by schoen 11y ago
One important element would be to rotate the hardware identifiers frequently, and, if the network is collecting payment from its users, totally decouple billing from any persistent identifier (for example, with blinded signatures). You also want to decouple user-oriented communication services from network connectivity, so your mobile operator shouldn't give you a phone number (or e-mail address), just generic IP connectivity to reach other services that give you those things. That's a start!
The reason that we have persist identifiers with mobile phones historically is all about billing and call completion. If we did the billing and call completion at other layers that weren't tied to a mobile subscriber's identity, we could ditch the persistent identifiers and use much more temporary ones (much like you could still get on a wifi network with DHCP even if you changed your MAC address every 5 minutes, and you could also netcat anonymous voucher tokens to a network provider's server to prove your entitlement to receive connectivity).
- superuser2 11y agoReceiving calls is a critical (/the only) point of a cell phone for many people. How do you route the right call to the right hardware if you can't associate the hardware with a phone number?
- schoen 11y agoIf you want to do that, you need to register with someone who can give you a phone number, but ideally not your carrier. You could conceivably have a SIP-like model where you can send control connections over Tor and media connections over the regular Internet. (I think the Guardian Project has tried making actual VoIP calls over TCP over Tor, and I don't know how well it worked out for them. So I'm assuming you won't be able to route the media streams anonymously, maybe just the control channel.)