Slashdot Mirror
IETF and wiretapping standards
Anonymous Coward writes "I just noticed that the IETF has sent out a request for discussion dealing with the implementation of wiretapping in Internet Protocols. The motivation is based on laws some Governments have about telecommunication systems." The message and subscription information to their discussion email list, punningly titled "Raven", are available on the web. Oh, and "some Governments" includes the U.S. and most other countries, so I hope the IETF will get some good feedback.
The mailing list is public. You can subscribe here and read the archives here. This, IMHO, is good. The existing posts on the list are, for the most part, high quality, constructive and thoughtful. One would hope that this being posted to Slashdot doesn't change that.
The second reason it's uneconomical is because it's alot easier to place a hardware bug into current systems (plug in a system board, replace the network card with a lookalike and a transmitter, tempest, etc) than to tap the upstream site(s) they will be using.
The third and final reason it's uneconomical is because this all assumes the would-be criminal isn't using encryption - and if he's savvy he likely is. So what's the point? They wouldn't be able to spy on the criminals anyway - just the average american who thinks IE and outlook express are the greatest programs ever.
Soooo... my take on it? The feds want to monitor domestic communications, because anything else is impractical - too expensive even for the Big Three.
--
I have to disagree with this statement. If you pay any attention to the crypto world, especially lately with the US gov't trying to find a new standard for encryption to replace DES and all its associated conversations, you should understand that the reason encryption algorithms become "popular" and "standard" is because they are subjected to brutal levels of scrutiny and analysis to determine their ability to withstand the various attacks to which you can subject crypto algorithms. The ones that stand up the best to this sort of hammering are the ones that tend to become widely used simply because they can stand up to the worst sorts of attacks the smartest people in crypto can come up with.
Saying "using a non-standard algorithm is more secure than a standard one" is just as bad as saying "security through obscurity works." It might, but then again it might not. The whole point is that you just don't know, while with the routines that have been publically anaylzed, you do know, at least to a reasonable measure.
And as far as what sort of computing power the Governments might have (The U.S. and Japan in particular since they seem to produce the largest number of the most powerful supercomputers), there's a lot of "scare" noise being thrown about that I personally don't put much faith in. Most of the crypto algorithms are such that it would take a dramatic mathematical breakthrough to really crack them rather than just more horsepower. If it's the difference between not cracking a code before the heat death of the universe and getting into it just after the sun collapses into a brown dwarf, i'm not going to worry.
-=-=-=-=-
-=-=-=-=-
My mom's going to kick you in the face!
I wouldn't be so quick to use PGP, or GPG, in such a circumstance. PGP headers, sure, but it makes no sense to give the cracking team clues as to how the message might be broken.
If you want real security, use a non-standard algorithm (Skipjack or Rijndael are good for this) to encrypt the message. Put a fake key and PGP headers round it, and finally run it through a steganography package.
If the message is found, the chances are that they'd attack the key. If they broke the code and got the key, they're no better off, as it won't work. (If you're =really= clever, reverse engineer a key that =appears= to work, generating a valid, but meaningless, message when applied.) Alternatively, they might try to attack the message itself. No good, as you're not using any of the algorithms the package you claim to be using has.
It (almost) doesn't matter how good the actual algorithm is, if you can convince potential attackers that you're using something else entirely. You only need to be concerned if they discover the deception and fathom out (somehow) what system you really have used. Even then, you aren't entirely vulnerable. A strong algorithm is going to take a long time to break, and there are plenty of twists you can add. (eg: Store the message backwards, or swap adjacent letters, to try and fool algorithms for detecting possible keys into recording a false negative.)
The problem is that Governments don't NEED to care about encryption. They've ultra-powerful computers capable of feats that would blow the socks off Seymore Cray if he were still alive. (Mind you, if ghosts need socks, they still might!) So long as the Governments can get the raw packets, they're home and dry. Almost. They use computers to break codes. Computers are fast, but notoriously stupid. An ingenious cryptographer should be able to deceive even the fastest, most powerful code-breaking computers in the world to report false positives. Do that, and trick the operators into using the wrong decryption algorithms, you have some limited influence over what those operators see.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
then they should do it themselves, instead of mandating that everyone make it "easy" for them. Especially here in the United States of Amerika, nothing says I need to make it easy for anyone to understand what I'm saying.
Not to mention that people who really want to have private conversations still will be able to, by piggybacking on top of (or tunneling with) "truly" secure protocols. There are internet phone apps that use PGP, will probably ones that use GPG, there are secure ytalk's floating around, etc.
The hell with government observation. It's their problem if they can't read my mail, or tap my phone, not mine, nor my ISP's.
The IETF, bowing down to opening up holes in secure protocols, will IMHO, completely invalidate any stance they have about any commitment to security. After all, would you buy a safe which is secure, "except for this spot right here, which will only be cut through by Authorized Personnel [or anyone else who tries]"?
Compromising security for the sole purpose of being friendly to government is ridiculous. Do you think they'd reciprocate on their own security so that we can tap into their communications? Of course not. But then, who ever said life was fair?
Even compromising security so that something will be accepted for use in multiple countries doesn't work. What self-respecting nation would want to use something that has backdoors the US (or any other) government can use to eavesdrop on its citizenry? Even when told, "We won't do it unless we have to. We mean it this time. You can trust us. Would we lie? Again?" I seriously doubt anyone with even a modicum of concern would believe that, or use a backdoored protocol.
Just look at the Clipper chip, the export version of Lotus Notes, etc. How many do you see in widespread deployment?
My personal feelings are echoed by the statement (Jefferson?) that people who choose to give up some freedom for security deserve neither. And yes, I would rather see a criminal get away with a crime to avoid sacrificing any innocent's privacy, since only dumb criminals would use dumb protocols to begin with.
"The urge to save humanity is almost always a false front for the urge to rule." --H.L. Mencken
We are non-profit, grass-roots, and in the crucial early stages of development.
Our goal is to develop a publically available VPN based on IPv6 and IPSec. We hope to be a public domain for serving 21st Cent. things likes VoIP, application servers, anonymizing proxies. We also seek to make cheap computers and free (speech) software available to low income families and individuals.
I invite you to see www.ompages.com. If privacy is an issue for you and you want to do more than 'write your local congressman', for example, by donating skills, equipment and resources to the public works project to build a secure network then join us. There is no leader, you can start your own project on ompages that furthers our goals of private networks and global technology proliferation. There will be no public network where individual privacy rights are the prime goal unless intelligent and experienced sys admins, programmers and web-masters get on the ball and make it happen. Talk is cheap; we can do this.
We must speak with one international voice against privacy intrusions to the IETF. If the IETF won't give us the privacy protections that are our birth rights, then we must implement our own. In fact, AOL users should not be subjected to the hoodwinking they are receiving. It is our duty as technically educated net citizens to give them the services they have now in a much more secure environment. Our priority is not the bottom-line; it's the line that must be maintained between individuality and government sponsored controls. This is no small task, but then again, neither is freedom. The U.S. claims to be governed by the people; ompages.com is.
My though is that putting wiretap capabilities into the lowest levels of the protocols is useless. So you can tap the IPv6 packet layer. So what? I'll just use SSL above that, or PGP-encrypt my mail, and your tap is useless.
There's also this: countries feel they need the Internet. Perhaps it's time to use the leverage this gives. Make no allowances in the protocols for wiretapping and the like, and give various countries a choice: allow people their privacy, or you will not be able to interoperate with the Internet. As noted above there are too many ways the people the governments could legitimately tap can bypass any hooks in the protocols, and why should the Internet protocols be designed to even potentially compromise the privacy of those who aren't legitimate targets?
That way, if a company wants to implement and sell a product that meets the standard in a way that fascistic governments who don't believe in personal freedoms will let them build and sell them, they can do so by implementing the "OPTIONAL" Backdoor parts of the spec.
Those groups who prefer security over letting Uncle Sam (or whichever hacker group out there is simply smart enough to read the specs and implement their own snooping software that follows the "RFC-'1984' - Government Backdoors into Network Protocols" spec) from eavesdropping, like the OpenBSD guys, can simply ignore the "OPTIONAL" part of the spec that outlines the backdoor without breaking the entire thing.
Sorry for the emotionally-loaded phrasing, but this kind of crap really gets me steamed. I'm amazed on a daily basis at how willing our governments are (especially here in the US) to simply trample our civil/constitutional rights for the Holy purpose of "National Security" whatever that means.
-=-=-=-=-
-=-=-=-=-
My mom's going to kick you in the face!