Profile

unixronin: Galen the technomage, from Babylon 5: Crusade (Default)
Unixronin

December 2012

S M T W T F S
      1
2345678
9101112131415
16171819202122
23242526272829
3031     

Most Popular Tags

Expand Cut Tags

No cut tags
Sunday, October 11th, 2009 05:42 pm

Google's Voice service lets you have a single master number that is automatically forwarded to as many other numbers as you designate, within reason. It's a good idea. But the US is running out of phone numbers as it is, and there probably aren't enough allocatable phone numbers remaining to let everyone do this without revamping the entire US phone number system.

But what if we rethought the whole concept of phone numbers?

Historically, one number has been tied to one phone line. You may have multiple phones on that line, and you may have a phone that has multiple lines coming into it, and ISDN phones have the capability to have up to three "appearances" of the same number on the line. But still, one number is tied to one line — or, in the case of mobile phones, one number is tied to one mobile phone. This metaphor has been a good match for the way the technology has always worked. But the technology is changing — already has changed enormously — and we don't have to do that any more.

Suppose we decoupled phone numbers from devices, and tied them instead to people and to roles, replacing a 1:1 mapping between phones and numbers with a many-to-many mapping. You have, for example, a personal phone number, an office phone number, and an on-call phone number. Your home phone is subscribed to your personal number — except that on days when you work from home, you subscribe it for the day to your office number as well. Your mobile phone is subscribed to your personal number and your on-call number. The phone on your office desk at work is subscribed to your office number and your on-call number. Perhaps you have a team number as well, that everyone on your work team is subscribed to. Whenever any of your phones rings, the display tells you not only who's calling you, but which of your subscribed numbers they're calling — and which of their numbers they're calling you from. That incoming call on your office phone is from your Cisco rep, but he's calling you from his personal number, perhaps to let you know he's in the area and ask whether you want to meet for lunch.

Discuss....

Tags:
Sunday, October 11th, 2009 11:37 pm (UTC)
How thorough would this profile have to be, and what would it do to identity theft?

I have a unique name. What would Terry Scott, a former coworker, need to have to make sure that anyone wanting to connect to her would actually get her?
Monday, October 12th, 2009 12:19 am (UTC)
I have a unique name. What would Terry Scott, a former coworker, need to have to make sure that anyone wanting to connect to her would actually get her?
What does she do now?

If, when you ask "what would it do to identity theft", you're thinking about someone else subscribing to your phone number and masquerading as you, that problem is one that can probably be best solved by using public-key cryptography to secure and authenticate subscriptions.
Monday, October 12th, 2009 12:21 am (UTC)
She's still there. I'm the one who's gone ;)

No, I'm not thinking about someone pretending to be someone else in that sense. I'm more concerned with having yet-another database with my info in it.
Monday, October 12th, 2009 12:55 am (UTC)
Why would it be any worse than the phone book?

Your phone company already stores far more information about you than your phone listing makes available to the public. The only additional piece of information they'd need to store would be a public key, which could be a dedicated key that you use ONLY for phone authentication, and which would do no identity thief any use unless they also possessed your matching secret key.

If anything, the need for that key to authenticate you for subscription purposes would strengthen your protection against identity theft. Cracking a 2048-bit public key by brute force is so computationally infeasible that, if you could capture all the rest of the Sun's output for the rest of its stellar lifecycle and dedicate all of that energy to the task, you could not exhaust the entire keyspace. Barring completely revolutionary new mathematical attacks, properly-constructed 2048-bit public-key encryption is, for all even remotely practical purposes, unbreakable.
Edited 2009-10-12 01:03 am (UTC)
Monday, October 12th, 2009 06:30 pm (UTC)

While I generally agree re: the intractability of NP–complete problems, there are a couple of important caveats.

  • PKI is sorely lacking and nobody really has any idea how to roll it out
  • There’s strong evidence (http://crypto.stanford.edu/~dabo/abstracts/no_rsa_red.html) that breaking rsa is not dependent on factoring large composites
  • It is thermodynamically possible to use the Sun’s energy to exhaust the space of 2048–bit primes. Even a 4k factoring challenge only has a work factor of around 2**168; the Sun’s lifetime energy is about enough to do a work factor of 2**210.

In the main, though, I concur.

Monday, October 12th, 2009 06:58 pm (UTC)
I guess my math was off. (Or more to the point, I think, I was remembering the math for the wrong problem.) Mea culpa.

I hadn't heard about that result...
Edited 2009-10-12 06:59 pm (UTC)