3 ms·

On spam filtering: the public telephone network transmits almost zero data on a call - the only header is the originating number, which can be un-traceably spoo

On spam filtering: the public telephone network transmits almost zero data on a call - the only header is the originating number, which can be un-traceably spoofed at the originator. There is no embedded message data (that would be the call itself). So there’s literally no valid data to filter spam on, unless the spam caller is chivalrous enough to use their legit caller ID. I would also bet that there are regulations that require phone providers to complete all calls that are presented to them. My provider, T Mobile, does overwrite the caller ID with “Scam Likely” for what they believe to be telemarketing calls, which is probably the best they can do until we implement a standard that authenticates the origination ID of calls.

bradleyjg 7y ago
There’s at least one piece of data on every call that the network operator has—-the identity of the prior network hop. In the early days of fighting email spam, that was enough. Every network operator was responsible for receiving spam complaints and responding appropriately. Network providers that couldn’t or wouldn’t prevent spammers from using their networks were blacklisted and other networks would refuse to route mail originating from them. It wasn’t perfect, but it was far better than nothing.
- rsync 7y ago
  "There’s at least one piece of data on every call that the network operator has—-the identity of the prior network hop." I'm curious - is this available in a call passed through twilio ? I would like to log, or pass, the prior network hop using twiml ... or a twilio function, I guess ...
- jgowdy 7y ago
  From my best interpretation of your comment you seem to think VoIP is far more "carrier peer to carrier peer" than it is in practice. As a value added telecom provider I have a few upstream carriers but it's not as if I receive calls from every carrier. Just like the Internet, my upstream carriers are my access points to the PSTN, so all of my SIP invites come from either my customers or my know upstream carriers. Those carriers are effectively my ISPs of the PSTN. I chose them and I pay them for service. Distrusting one of them and the calls that come from them is something of a non-sequitur since you chose them and intentionally became one of their customers and pay them to receive those calls. TL;DR This isn't like email or IP traffic where you're interacting with other IPs in a purely untrusted network model. Merely analyzing the previous hop doesn't tell us anything valuable.
  - bradleyjg 7y ago
    That’s the peril of being a leaf node. The same thing was true in the early days of email. An end user, or in your case something close to it, couldn’t blacklist his upstream—-as you say that would be ridiculous. But your upstream (or possibly their upstream) does have true peers that they could blacklist for bad behavior. It might mean there customers couldn’t receive calls from entire countries, but that’s a trade-off I, and I assume plenty of other, would be happy to make and even pay extra for. This assumes that is that there isn’t some kind of international treaty that forbids doing so. The existence of such has been claimed from time to time in similar discussions but I have yet to see an actual citation.
mschuster91 7y ago
> the only header is the originating number, which can be un-traceably spoofed at the originator That's wrong, phone signalling systems always specify the originating number and it cannot be spoofed. The key thing is that only 911 responders and similar lines have the CLIRIGN flag set on their lines which means that the spoofed number transmitted via CLIP is ignored and the real one used instead. In addition victims of e.g. threat calls can have the MCID flag set on their line which makes the provider systems log all call information for later retrieval by police.

bradleyjg 7y ago

There’s at least one piece of data on every call that the network operator has—-the identity of the prior network hop. In the early days of fighting email spam, that was enough. Every network operator was responsible for receiving spam complaints and responding appropriately. Network providers that couldn’t or wouldn’t prevent spammers from using their networks were blacklisted and other networks would refuse to route mail originating from them. It wasn’t perfect, but it was far better than nothing.

rsync 7y ago
"There’s at least one piece of data on every call that the network operator has—-the identity of the prior network hop." I'm curious - is this available in a call passed through twilio ? I would like to log, or pass, the prior network hop using twiml ... or a twilio function, I guess ...
jgowdy 7y ago
From my best interpretation of your comment you seem to think VoIP is far more "carrier peer to carrier peer" than it is in practice. As a value added telecom provider I have a few upstream carriers but it's not as if I receive calls from every carrier. Just like the Internet, my upstream carriers are my access points to the PSTN, so all of my SIP invites come from either my customers or my know upstream carriers. Those carriers are effectively my ISPs of the PSTN. I chose them and I pay them for service. Distrusting one of them and the calls that come from them is something of a non-sequitur since you chose them and intentionally became one of their customers and pay them to receive those calls. TL;DR This isn't like email or IP traffic where you're interacting with other IPs in a purely untrusted network model. Merely analyzing the previous hop doesn't tell us anything valuable.
- bradleyjg 7y ago
  That’s the peril of being a leaf node. The same thing was true in the early days of email. An end user, or in your case something close to it, couldn’t blacklist his upstream—-as you say that would be ridiculous. But your upstream (or possibly their upstream) does have true peers that they could blacklist for bad behavior. It might mean there customers couldn’t receive calls from entire countries, but that’s a trade-off I, and I assume plenty of other, would be happy to make and even pay extra for. This assumes that is that there isn’t some kind of international treaty that forbids doing so. The existence of such has been claimed from time to time in similar discussions but I have yet to see an actual citation.

mschuster91 7y ago

> the only header is the originating number, which can be un-traceably spoofed at the originator That's wrong, phone signalling systems always specify the originating number and it cannot be spoofed. The key thing is that only 911 responders and similar lines have the CLIRIGN flag set on their lines which means that the spoofed number transmitted via CLIP is ignored and the real one used instead. In addition victims of e.g. threat calls can have the MCID flag set on their line which makes the provider systems log all call information for later retrieval by police.