|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |
Date: Mon, 14 Apr 1997 21:42:31 -0700 From: Hal Finney <hal@rain.org> To: cryptography@c2.net Bill Frantz, frantz@netcom.com, writes: > Why do you assume that just because a CA has issued a cert for a public > key, that the corrisponding secret key holder had anything to do with it? To answer this as narrowly as possible, I would assume this because it would be true for probably 99.999% of the certificates issued by a respectable CA. In virtually all cases, a CA issues a certificate upon receiving a request from the key holder. The point of John's story, and of mine, was to remind people that this is not the only way things could happen. If a CA violates its own policies then if other people make the assumption above (which again is a plausible assumption in almost all cases), the CA can get powers similar to a key escrow agent. > There is nothing preventing a CA from issuing a cert for any key it can > download from a public key server. Or, as in my example, a key that it makes up all by itself. > The statements from a CA flow from the CA to the key, not in reverse. The problem arises when the CA lies. When it says, "this is the key to use for secure communication to www.ibm.com", and gives a false key that it knows the secret to, it can potentially read traffic to that web site. Hal
|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |