home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!uwm.edu!rutgers!network.ucsd.edu!qualcom.qualcomm.com!servo.qualcomm.com!karn
- From: karn@servo.qualcomm.com (Phil Karn)
- Newsgroups: sci.crypt
- Subject: Re: Ciphers
- Message-ID: <1993Jan4.110356.23995@qualcomm.com>
- Date: 4 Jan 93 11:03:56 GMT
- References: <R90TwB1w165w@west.darkside.com>
- Sender: news@qualcomm.com
- Organization: Qualcomm, Inc
- Lines: 46
- Nntp-Posting-Host: servo.qualcomm.com
-
- In article <R90TwB1w165w@west.darkside.com> sspy@west.darkside.com (mshannon) writes:
- >Another question, gentlemen. Suppose someone, the feds or whoever, hace
- >an encrypted disk that they want to crack, but they don't know what
- >cipher was used. How can they find out. Are there programs that will
- >analyze it and make this determination based on the way certain ciphers
- >cipher?
-
- In general, no. By itself, any good cipher produces ciphertext that is
- indistinguishable from purely random bits by anyone who doesn't know
- the key. In practice, however, many encryption packages format their
- ciphertext output to make it easy for the intended recipient to
- recognize as a message encrypted with that particular package. This
- will tip off an eavesdropper that a particular system and possibly a
- particular set of keys are in use (implying the identities of the
- correspondents), even if he can't actually crack the contents.
-
- For example, PGP puts out the "BEGIN PGP MESSAGE" banner, and it
- specifies the public key that was used to encrypt the message so the
- receiving program knows which of possibly several secret keys to use
- to decrypt it. Anyone with a reasonably complete public key collection
- can see who an intercepted message is addressed to, at least at the
- outermost level (it might contain several levels of encapsulated
- encryption, as in some of the anonymous remailer schemes).
-
- Other encryption packages, such as Sun's 'des' command (and the one I
- wrote that's compatible with it) produce *no* identifying information
- in the ciphertext -- it looks like purely random bits. This has the
- advantage of denying an eavesdropper knowledge of the system you're
- using, but it has the disadvantage that the user must know, a priori,
- which algorithm and key to use to decrypt it. This is okay in a single
- key cryptosystem (you could consider the algorithm specifier to be
- part of the "key" that you exchange ahead of time) but it is not
- practical in a public key cryptosystem, which is supposed to let you
- send encrypted messages to someone without special prior arrangement.
-
- On the other hand, it would not be hard to take that little bogus PGP
- ciphertext generator I posted the other week and go much farther with
- it. It could produce a properly formatted PGP message, with correct
- checksums, etc, that causes PGP on the receiving end to prompt for a
- totally random and fictictious key. To an eavesdropper, such messages
- would be indistinguishable from real PGP messages encrypted with
- unknown public keys. One could conceive of generating continuous
- streams of such traffic just to provide some haystacks in which to
- hide your needles.
-
- Phil
-