[nadams]

No offense but I didn't mention a specific VOIP provider on purpose.

However, callcentric already has it built and much easier to understand than twilio. I'm not saying that service is bad - but the average person doesn't know what API, cloud, or SIP means (or care). Callcentric makes it pretty obvious what their service is and provides and how to use it.

[toomuchtodo]

I specifically mentioned Twilio because anyone can build a custom solution to address this problem.

Someone already has done this (with Twilio) to win the FTCs RoboCall Challenge: https://www.twilio.com/blog/2013/04/twilio-powered-nomorobo-...

I also like Twilio ;)

[MichaelGG]

Funny. I've handled a lot of dialer traffic for customers, as an intermediary (a billion calls every day or two).

"Legal" robo dialer is fairly easy to block. We blocked it on our systems using the same ideas as in the link. Look for repeated calls, low call duration, etc. However, this did not impact our calling stats at all.

That's right, even filtering out all the "known" dialer, we still had tons of shitty dialer traffic coming. The scammier the dialer, the more inclined they are to just make up a new caller ID for each call. Then there's no real pattern and it's easy to slip through. This is illegal and the FCC can fine $10K a day or something. But they don't.

That's the REAL problem. There's never any follow up. Sure, end-users complain now and then. What does that do? It makes one telco pass up a request "please review or block this number". That's it. It is utterly ineffective.

If the FCC was actually serious, they'd start enforcing rules and hold companies accountable. If I, as a VoIP provider, am liable for my customers to a certain extent, then I'm gonna be a bit more selective. I'll make sure my $29 customers aren't faking caller ID or sending lots of calls. On the wholesale side, I will make sure my contracts contain harsh penalties and keep customers liable. It'll take a few months for stuff to trickle down and be enforceable, but eventually the FCC can simply start handing out fines and this whole game will be over.

There's nothing really stopping this from happening other than the FCC not wanting to do it. Sometimes people say "oh well the problem is someone in Elbonia can sign up for $15 and get going" - yeah well eventually their calls hit the US network, and US law applies. If the provider is letting the Elbonians do their scam, the provider is liable. End of story. Suddenly, providers will require a deposit and scammers will piss off.

It doesn't even have to be a large fine. Just making the cost of making up a new identity be non-trivial (a few hundred dollars is probably enough) will put a pin in it.

[nadams]

> There are no infrastructure changes, or hardware required. This was one of FTC’s guidelines for the contest.

> Unfortunately, Nomorobo is not available on traditional analog landlines or wireless phones at this time.

Sounds like he failed the challenge...

I don't think his idea is revolutionary - one could setup Asterisk to do what he did.

[patio11]

One is welcome to ship an Asterisk implementation of it any time one feels like doing so. One has not shipped an Asterisk implementation. One may underestimate to which degree ideas, revolutionary or otherwise, are improved by actually shipping concrete instantiations of them which actually work.

Incidentally: NoMoRobo gains some utility specifically because Twilio exists. A trivial example: phone numbers have a history to them. Most customers of Twilio want to buy phone numbers with a relatively clean history, i.e., ones that were not widely distributed prior to their life with that customer, to avoid misdirected calls or e.g. reputational issues. NoMoRobo wants precisely the opposite -- they literally ask Twilio to provision them with phone numbers which are otherwise unsaleable to customers, for example because those numbers are actively getting spammed to death. (Twilio has an entire team of people whose job is telephone number quality. Things you wouldn't have guessed existed in the world, right?)

This works great for NoMoRobo's use case, because a phone number getting spammed to death is perfect for them -- it lets them cheaply capture a lot of phone numbers which one has a high confidence are spam rather than ham.

[MichaelGG]

Fun/rude anecdote: We (a VoIP provider) once got a number that was getting 250K calls every week. Not because telemarketers were calling it, but because that ID was being usd to make a lot of outbound calls. So people would call back, upset, trying to figure stuff out. We dumped the number (no sense paying for all that traffic), but not before shunting it into a conference call (and playing a message telling them so).

It was rather interesting to see how people reacted together. One person would call in, angry, another person would call in, soon there'd be 4 people all yelling at each other "Well you called me!" "No I didn't!" "Now listen here, I think maybe the wires are crossed" and on and on. Sometimes they'd gang up together and try to sort problems out, but often it'd just degenerate into name calling.

Anyways, it's sad that NoMoRobo exists. The FCC could trivially put an end to scam calls, robo-dialing, etc., with a couple days of drafting a new regulation, and then a few months to implement it. Source: I've handled billions of calls, much of which were dialer traffic. I've worked on both sides (trying to block traffic, trying to get upstreams to take traffic).

[toomuchtodo]

> with a couple days of drafting a new regulation, and then a few months to implement it.

How would they enforce the regulations though without NSA style SS7 taps throughout US telcom network infrastructure?

[plonh]

Why would you intentionally harass victims like that?

[toomuchtodo]

Wow, thanks so much for contributing this Patrick. I had no idea Twilio did this or had a team who worked on it!

[toomuchtodo]

> I don't think his idea is revolutionary - one could setup Asterisk to do what he did.

Twilio is stupid simple to prototype and get up and running with, at which point you could transition to your own SIP provider with a cluster of Asterisk servers when you're at scale.