Hi,

I'm doing some small development (crytpo protocol, similiar to ssl but
by far not that complex) with the openssl/libcrypto routines.

Here is a short list of comments and questions could not been answered
using the the manual pages or the sources (or google :).

- What is the risk in using the default OpenSSL 
  DH parameters? What is the benefit in generating them?
  (generating 4096b DH parameters on a 1.2Ghz box takes some hours/days :/).
  Is a /dev/urandom (linux 2.4) source to seed the
  OpenSSL PRNG fine? (seen it in many documentations,
  still dont feel very comfortable about it).

- The DH shared secret is to long for most symmetric
  ciphers. I'm using EVP_BytesToKey() to generate
  suiteable key + iv for the symmetric cipher.

- Both parties should use a different key + iv for both directions. Eve
  could otherwise figure out that both parties start the communication
  with the same words (same crypted packets => both parties say the same).
  I simply invert the shared shared secret before
  it is passed to EVP_BytesToKey() to generate two different
  key + iv pairs (one pair from the original shared secret, the
  other pair from the inverted shared secret).

- I use the MD5-sum of the shared secret as HMAC key.
  The HMAC key is published _after_ the communication
  finished. Any problems with this?

- Any way to determine how many % of the DH parameter generation
  has finsihed? (Pretty hard to predict when a prime 'will be found'
  but dont blame me for asking :)


Thanks for brainstorming with me about these rather simple questions :)


skyper
-- 
PGP: dig @segfault.net skyper axfr|grep TX|cut -f2 -d\"|sort|cut -f2 -d\;
______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [EMAIL PROTECTED]
Automated List Manager                           [EMAIL PROTECTED]

Reply via email to