[lucasjung]

One thing the article didn't mention, which I've been considering in regards to this problem: an RF-based internet alternative would be prone to all sorts of other forms of government interference/monitoring. The U.S. government already has serious resources at its disposal for the purpose of intercepting or jamming RF transmissions. For "intercepting," this includes high-power decryption capabilities, and for "jamming" this includes noise jamming but also spoofing and signal insertion. So even if an RF mesh-network of some sort were to be established, the government would be able to:

1: Know exactly where every transmitter is. This means they can find you in meatspace even more easily than they can on the hard-wired internet.

2: Listen in on your transmissions without all of the legal issues associated with wiretapping. To make sure they can do so, they would probably need to pass a law prohibiting the use of many types of cryptography on unlicensed RF transmissions. Such a law would be much easier to sell to the general public "because the terrorists could be using it to coordinate attacks." If you break this law, expect a knock on your door almost instantly because of #1, above.

When you combine #1 and #2, busting "pirates" becomes trivially easy: somebody sees a "suspicious" file in your transmissions, localizes your transmitter, and a few minutes later you get a knock on your door.

There's other stuff, like injecting false traffic, etc.

[IgorPartola]

Just to clarify: are you saying with your point #2 that the government can break any encryption currently in wide use with 802.11 networks and all popular VPN solutions? Or are you just saying that they can physically listen to encrypted transmissions? I was under the impression that encryption scales in complexity pretty much infinitely, so long as you don't care about encryption/decryption speeds.

[delinka]

I had a relative that once worked in "government security" for the U.S. government. What that meant, we weren't permitted to know. Personally, I think he was full of crap and here's why: in his Professional Opinion, any encryption algorithm implemented in software was doomed to failure because "it could be hacked" - he was completely unaware that any hardware circuit can be emulated in software.

Point: it tends to be these very gov't lackeys that think just because the signal is in the air that it can be intercepted, decoded, decrypted and its plaintext content recorded.

[lucasjung]

I'm saying that any government (possibly not some of the more representative governments, but certainly many of the more oppressive ones) could outlaw for unlicensed RF transmitters any encryption they are incapable of breaking.

As to what NSA, et al are currently capable of, I honestly have no idea, but I'm willing to bet some of their capabilities would be surprising (in both directions, depending on which specific capability you were to look at).

[nate_meurer]

This isn't what the parent was saying, but you are correct: the NSA can't break properly implemented strong encryption. They can't perform miracles; they used the same technology the rest of us use, just on a different scale.

Fifteen years ago we had proof of this after the government inadvertently showed its cards via clipper chip, the crusade against PGP, etc. We still have export controls on strong encryption, and there's a good reason for that (well, maybe not a good reason).

[shaddi]

I'm less worried about this actually -- I think the biggest problem is your own nodes interfering with themselves ("internal interference") rather than with an attacker. Agreed this is an issue though, and using omnidirectional antennas exacerbates it. Using directional antennas actually helps, for the same reason it's hard to see a laser beam from the side.