Re: Cracking decrypted file when knowing partial contents

From: Bill Unruh (unruh_at_string.physics.ubc.ca)
Date: 04/20/04 

Date: Tue, 20 Apr 2004 17:48:02 +0000 (UTC)

"Adfa" <adfa@noname.com> writes:

]Would it be easier to crack an encrypted file if the cracker knew some of
]the contents of the file?

]For example, an encrypted MS Word file where the cracker knew word-for-word
]a few paragraphs of the document, but not necessarily the exact byte
]postions of those paragraphs in the file.

It depends on the encryption used. Most modern good encryption is
designed to resist known plaintext attacks-- ie even if you know the
plaintext and how it encrypts, you still cannot determine the key except
by exhaustive search of the key space.

Known plaintext attacks are so well known (That was how individual
enigma messages were cracked) that any competent cryptographer spends
most of his time designing the encryption scheme to be resistant,
although even compentent ones can screw up.

Of course even if designed to resist known plaintext attacks, the key
space may be so small that exhaustive search has a chance of
succeeding.

HOwever it is easy to design encryption schemes which do not satisfy the
above criterion of a good encryption scheme. Most home grown schemes -- whether
by some backwoodman encrypting his diary, or Microsoft or some other
software company including "encryption" with their products--
are bad. Only use well known schemes which have had lots of scrutiny by
the community. Designing your own, either because you don't trust the
public schemes, or because you are a genius, is like trusting yourself
to fly in a plane designed by amatuers, and piloted by an untrained
pilot.

Relevant Pages

  • Re: Telephone Scramblers
    ... If one were to set about designing such a ... VoIP will make various digital encryption techniques easy to implement. ...
    (sci.electronics.design)
  • Re: Cracking decrypted file when knowing partial contents
    ... an encrypted MS Word file where the cracker knew word-for-word ... It depends on the encryption used. ... above criterion of a good encryption scheme. ... Designing your own, either because you don't trust the ...
    (sci.crypt)
  • Re: Storing user information as RC4 - safe?
    ... Designing a cryptosystem is highly subtle stuff. ... I'm new to encryption, but not coding. ...
    (sci.crypt)
  • Re: Encrypt and Authenticate with same key
    ... The `security warranty' tells you under what circumstances you ... a one-time pad is a symmetric encryption scheme. ... properties against computationally unbounded adversaries. ...
    (sci.crypt)
  • Re: secure passwords
    ... >>Except that the whole idea of shadow passwords is that the cracker does ... >>not have direct access to the password hashes, so he must go through an ... able to crack a typical DES 8-character password in a matter of monghts, ... even if the robustness of the password encryption is *perfect*. ...
    (comp.os.linux.security)