So it's been all over the news that a judge ruled that the nsa wiretapping program is illegal etc etc etc. I'm not a lawyer, so I'm not really going to go into the legal aspect, however, I got to thinking from a technology perspective, if I was asked to create an effective program to track terror communications over communication lines, how would I go about it?
First, we have to define the scope of the problem:
1. We don't know who the terrorists are.
2. We don't know when and where they are calling from.
3. There are tens of thousands of international communications going on 24/7.
So based upon this problem scope, how would one handle not only identifying terrorists, but ensuring that if one was identified you could quickly and effectively "connect the dots" and identify co-conspirators etc? There are a few obvious things based on the above assumptions:
1. There is no way for any team of humans to manually tap calls.
2. There is no way for a team of humans to listen to every single call.
3. If you do identify a terror suspect, you not only want future records, but past ones as well possibly including the calls of people linked to them via number cross reference.
With these concepts in mind, I can only come up with one conclusion. The only way to effectively handle this type of monitoring is to log all calls, indexed by datetime and phone numbers of both parties. This is a huge amount of data, being that no one can go through it all, the purpose of the system would be whenever any agency raised a red flag on someone, you could pull their records as well as the records of anyone they're in contact with, and then you get the warrant to crack the records open.
Thus, I think this whole hysteria from some people about "warrantless wiretapping" is kind of a FUD argument. If the NSA is recording all of my calls, but never listening to them without probable cause and a warrant, it doesn't really hurt anyone, in addition it maintains the history so if a red flag gets raised, you can go back and figure out what's been going on historically as well as future calls, as well as follow the links by phone number, getting as many warrants as you need to at this point. Because there are tens of thousands of calls a day, it is unfeasible to expect to get warrants to monitor every call, and it is also unfeasible to just delete the ones you aren't listening to, since they may need to be cross referenced and analyzed some day. However, I do think that as soon as a flag is raised, they best be getting a judge and warrant involved.