Secure Key Distribution

Levison has been jetting around the globe recently and presenting his project at various conferences that focus on security. He also attended the Chaos Communication Congress in Berlin at the end of last year [8], where he talked about the motives behind his actions. In talking to Levison, one quickly gains the impression that the man knows what he's doing. This is evident in the details, which were often ignored in the use of GnuPG. As recently as December, for example, Richard Klafter and Eric Swanson conclusively demonstrated that 32-bit IDs for keys in GnuPG are quite easily predictable [9]. If you try, you can – in a very short time – generate a key set with thousands of keys that are mutually signed and have trustworthy-looking entries in the name and email fields. In other words, the procedure described here gives hackers the ability to generate keys whose IDs are identical to the IDs of "genuine" personalities. The problem doesn't affect the actual keys of those personalities; they are still as secure as ever.

The problem is that spoofed keys are only identifiable if users take the trouble to compare the fingerprints of the key that they know with the one that they use for encrypting a message. If they fail to do so, they are opening up the door to exploitation. This is all the more tragic when you consider that GnuPG has for years offered the option of using 64-bit key IDs. Virtually nobody uses this function, and very few GnuPG users are even aware of the problem or that a solution exists for it.

DIME has problems of this kind in its sights. Key distribution within the service will come with built-in security out of the box: The key servers used there will be sufficiently robust to prevent them being hijacked by malevolent services with a large budget. Additionally, the key management system will be integrated at such a low level in Dark Mail that users won't even notice it exists.

DIME Has Potential and Needs to Deliver

Not a lot of information is available on DIME. Conceivably, Levison and his partners do not want to expose DIME too much before it can actually do what it promises. Secret services like the NSA will no doubt view DIME as an insult and do everything they can to prevent the service from being used. DIME is obviously interested in not giving them an attack point for doing so. To damage DIME in its currently known design, such agencies would effectively have to prohibit encrypted communications across the board. This appears unlikely right now but not totally impossible; cryptography software was, for a long time, handled like a weapon in the United States and subjected to strict controls, especially when exporting it to foreign countries.

Some comprehensive testing will be needed to discover what DIME really can do after it goes live. If the developers succeed in successfully launching the service and publishing the required software as open source, the system could really take off. One thing is certain, though, legacy email has reached the end of the road in terms of IT security.