5 ms·
The Spy in the Middle
- rmorrison 17y agoThey found turnkey surveillance products, marketed and sold to law enforcement and intelligence agencies in the US and foreign countries, designed to collect encrypted SSL traffic based on forged "look-alike" certificates obtained from cooperative certificate authorities. The products (apparently available only to government agencies) appear sophisticated, mature, and mass-produced, suggesting that "certified man-in-the-middle" web surveillance is at least commonplace and widespread enough to support an active vendor community. I wonder if China, Sudan, or Iran are included on the list of governments that they're allowed or willing to sell to. If so, the problems that arise from trusting too many Certificate Authorities would also apply to trusting too many governments.
- dkimball 17y agoWith the SSL-subversion boxes mentioned in http://www.wired.com/threatlevel/2010/03/packet-forensics/ http://www.wired.com/threatlevel/2010/03/packet-forensics/ , I don't think it matters who's on the list (though I'm sure China is). A small company selling devices to intercept SSL means that the whole concept is dead. Even if no one else can acquire their devices, the whole world (that cares to find it out) now knows that such a man-in-the-middle attack is possible, so it's just a matter of time before they become much more common. Worse yet, the company that makes these boxes advertised them at "the world's largest gathering of North American, Caribbean and Latin American Law Enforcement, Intelligence and Homeland Security Analysts and Telecom Operators responsible for lawful interception, electronic investigations and network Intelligence gathering." Any Mexican drug lord worth his cocaine money (or, of course, FARC) will now start bribing or scaring some suitable official into handing over a box or two. Hopefully we'll discover an alternative approach to ensure online security. Just having fewer, more reliable CAs would be enough -- although, as I pointed out elsewhere, the measure of security is ultimately boots on the ground, not code in the aether. Update: You're right, there's a trusted Chinese CA. From the Wired article: "[w]hen Mozilla added a Chinese company, China Internet Network Information Center, as a trusted Certificate Authority in Firefox this year, it set off a firestorm of debate, sparked by concerns that the Chinese government could convince the company to issue fake certificates to aid government surveillance."
- qjz 17y agoThe number of CAs doesn't matter. The essential flaw in SSL is that once you trust a CA, you trust it for all sites. This is what makes a MITM attack so trivial for someone who is well positioned in a network, and these appliances just make it that much easier to deploy if you don't have all the skills yourself. Another problem is the sheer conceptual complexity of the system. It wouldn't take much to convince your boss to let you purchase an "SSL proxy" to fight viruses and commercial espionage, when all you really want is to steal all his passwords.
- borism 17y agoUS has free trade with China. Sudan and Iran are "axis of evil" states anyway. China has close trade relations with both. No government can be blindly trusted no matter which one it is!
- wmf 17y agoWe've known that this is possible for years (http://crypto.stanford.edu/ssl-mitm/ http://crypto.stanford.edu/ssl-mitm/ http://www.darknet.org.uk/2009/08/sslsniff-v0-6-released-ssl-mitm-tool/ http://www.darknet.org.uk/2009/08/sslsniff-v0-6-released-ssl...), but the fact that it's now a commercial product indicates that it is actually happening for real.
- peterwwillis 17y agoI think an even better question is: what does it take to become a real CA? Don't just assume 'oh, it must be really tough' if you don't know for sure. How much money and what kind of resources would it take some to really create their own valid CA just for generating forged yet valid certs that browsers would not flag as invalid when used in a MITM such as those used by the device in the wired article?
- wmf 17y agoIt is fairly difficult: http://www.mozilla.org/projects/security/certs/policy/ http://www.mozilla.org/projects/security/certs/policy/
- peterwwillis 17y agoi'm sorry, but, how is that difficult? most of that involves basic procedures for verifying someone's identity before providing a cert plus an independent party for auditing and other purposes. you don't even need any money, just to set up the basic infrastructure and steps for issuing certs as documented in provided links. i think they basically hinge on the idea that they will match up your CA with real human beings and verify you gen certs safely, but this is not a roadblock for an attacker. i think with time, patience, and an apparent genuine interest in doing things the right way, anyone could get mozilla to accept their root certificate. (and then generate certs for the black market for unauthorized people with loads of cash)
- wmf 17y agoCAcert has been trying to get into Mozilla for years and failing: http://wiki.cacert.org/InclusionStatus http://wiki.cacert.org/InclusionStatus
- olefoo 17y agoYou can create your own CA in 5 minutes with a linux box and a copy of the OpenSSL binary. The hard part is getting your signing certificate accepted into client programs, web browser manufacturers and operating system vendors are picky about who gets in and who doesn't. If you're willing to handle all of the key management yourself and you are only accepting certificates that you know you control... and you know for a fact that nobody snooped your private key from the signing station... and you never stored the signing key on any persistent storage in the clear... and you implicitly trust all connecting clients to not be spoofing authentication to users, and, and... Then you can say truthfully that you think you can trust certificates signed by that key maybe.