0
$\begingroup$

I've recently started reading the book "Applied Cryptography" written by Bruce Schneier.

Quite a few times I stumbled across protocols who send the identity outside of the encrypted message (see step 4).

Is there a certain reason to do so? What does speak against sending the identity within the encrypted message?

enter image description here

$\endgroup$
  • $\begingroup$ I don't see any especial reason except that sending encrypted identities will be inefficient, cost of applying encryption function, when identities are not necessarily to be confidential. Which is evident in step 1. $\endgroup$ – user110219 Mar 17 '15 at 12:13
  • $\begingroup$ You need to know what public key to use to authenticate it. Without an identity, you have a random encrypted blob. bobs key cannot be looked up unless you know it is Bob you are interacting with. $\endgroup$ – John Meacham Mar 18 '15 at 7:19
  • $\begingroup$ Although this might make perfectly sense in other scenarios, this argument doesn't seem to be valid for this example. Until now I do like the notion about the additional overhead for encryption the most. $\endgroup$ – HansMusterWhatElse Mar 18 '15 at 9:07
0
$\begingroup$

You can of course add a trusted path requirement to Trent (the trusted key distribution center) on the top of the given protocol. This is however not a requirement for the scheme itself.

Trent, as trusted third party serving multiple users, of course needs to know who to communicate to. So the identity cannot be encrypted with the public key of Alice or Bob.

As you can see the information between Alice and Bob is encrypted. This is no problem as both just use their own key to decrypt the messages between them.


That all said, there seem to be multiple attacks on the scheme - this seems to be one of them. The problem of not establishing a trusted path to the KDC could be a big improvement on the protocol. Just to remind you: only use this protocol for study, not for real life implementations.

Please understand that these protocols are building blocks for creating secure systems. You need to understand exactly what kind of security is provided by the scheme to use it for building a secure system. Anonimity of Alice and Bob when talking to the KDC is clearly not part of the security provided by this protocol.

| improve this answer | |
$\endgroup$
  • $\begingroup$ "Trent, as trusted third party serving multiple users, of course needs to know who to communicate to. So the identity cannot be encrypted with the public key of Alice or Bob." - In step four the public key of Trent is used for the encryption, so that shouldn't constitute a problem. Your other two arguments about the scheme and the additional overhead make perfectly sense to me. This book is definitely just for studying and getting the whole concept of hashing / encryption. $\endgroup$ – HansMusterWhatElse Mar 17 '15 at 16:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.