Secure Mail at LCS

General Information

The Secure Mail project hopes to establish a secure e-mail infrastructure for LCS. Why don't people at LCS use secure e-mail? Would a secure e-mail infrastructure enable people to get things done that they currently can not?

We wish to transparently provide end-to-end confidentiality and authentication for e-mail. The recipient must take responsibility for storing mail securely after it has been read. The key requirements are usability and transparency. These are probably the reason why secure mail has not been more widely deployed.

Also, let's not forget Why Cryptography is Harder Than it Looks (Bruce Schneier).

Plan?

• Evaluate standards for securing e-mail.
  • Widespread use outside of MIT/LCS?
  • Integration into common mail clients? (Unix, Windows, Macintosh)
  • Flexibility of key management/distribution infrastructure?
• Select a standard for use at LCS.
• Implement standard's key management. (Alternately, make our own, e.g. like MIT's old PGP KeySigner; see the paper by Schiller and Atkins below.)
• Find or implement mailing list support.
• Produce documentation showing how to:
  • Obtain software,
  • Integrate software into user's existing mail configuration,
  • Generate and manage keys.
  • Use software on day-to-day basis.

Much of the software needed has probably already been implemented so the task is mainly to find, categorize, document and deploy existing software.

Resources

Standards

• OpenPGP
  • multipart/encrypted and PGP/MIME
  • application/pgp
• S/MIME
• S/MIME and OpenPGP

Implementations

• PGP support in mail clients.
• mutt (PGP, in stable, and S/MIME, in 1.5 series).
• Ximian's Evolution (PGP, and S/MIME is on wish-list).
• Emacs Mailcrypt (PGP). Integration for Gnus, R-MAIL, MH-Rmail, etc.
• Integration with PINE (PGP).
• Raph Levien's premail.
• Netscape (S/MIME).
• OpenSSL toolkit, useful for S/MIME..

Papers

• J. Schiller and D. Atkins. Scaling the Web of Trust: Combining Kerberos and PGP to Provide Large Scale Authentication [PS]. Winter USENIX 1995.
• A. Whitten and J.D. Tygar. Why Johnny Can't Encrypt?. Proceedings of USENIX Security 1999.
• N. Borisov, I. Goldberg, E. Brewer. Off-the-record Communication, or Why Not to use PGP. Submitted to USENIX Security 2002.

Other projects

• Notes from Brad Templeton on achieving zero-user interface.
• Herbivore.
• PPS