[jp] Tips for encrypting strategies?

Jochen Hoenicke hoenicke at gmail.com
Thu Nov 23 13:34:35 EST 2006


2006/11/22, Ralph Alvy <ralph at ralphalvy.com>:
> I use Keyring and Note Studio for my encrypted notes that have critical
> data. Note Studio is supposed to have quite strong encryption:
>
> "Note Studio uses a form of encryption called 3-Way Encryption. This is
> a very secure encryption. There is no known, practical way to crack this
> encryption. A mathematical method of breaking 3-way encryption has been
> proposed, but this method required a bank of super computers and a long
> time (years)."
>
> I've always wondered how strong Keyring encryption is. Anyone know?

As a co-author of Keyring I know it, of course.  The details are on the homepage
http://gnukeyring.sf.net/crypto.html

In short, keyring-1.2.3 uses triple des (to be more exact 2-TDES in
ECB mode).  It has a key length of 112-bits. According to wikipedia
there are "certain chosen-plaintext or known-plaintext attacks and
thus it is officially designated to have only 80-bits of security". It
is probably not feasible, as it requires too much known-plaintext,
though.

3-way has 96-bits key-length.  There is a related key cryptanalysis,
but this is probably not feasible for this application.

Both methods cannot be broken even by the current computing power
(e.g. distributed.net) within the next decade (even considering
Moore's Law).

As always, it is not as simple as this.  You also have to consider how
the encryption key is generated from password, which encryption mode
you use (block ciphers always have ECB,CBC,OFB and CFB modes), and how
the IV is generated. Also the security provided by the encryption
cannot be better than your password. Furthermore, data may be leaked
by operating system or by a trojan application, there are side-channel
attacks, e.g. if you use keyring very much, some characters may be
permanently scratched into the Graffiti area.

The keyring-2.0 pre-release has stronger encryption using full 3-TDES
or AES with 128-256 bits (at user's choice) and better encryption key
generation.  However, this release is not supported by jpilot, yet.

Regards,
  Jochen



More information about the Jpilot mailing list