Lightweight email signatures (extended abstract)

  • Authors:
  • Ben Adida;David Chau;Susan Hohenberger;Ronald L. Rivest

  • Affiliations:
  • CSAIL, Massachusetts Institute of Technology, Cambridge, MA;CSAIL, Massachusetts Institute of Technology, Cambridge, MA;Zurich Research Laboratory, IBM Research, Rüschlikon;CSAIL, Massachusetts Institute of Technology, Cambridge, MA

  • Venue:
  • SCN'06 Proceedings of the 5th international conference on Security and Cryptography for Networks
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present Lightweight Email Signatures (LES), a simple cryptographic architecture for authenticating email. LES is an extension of DKIM, the recent IETF effort to standardize domain-based email signatures. LES shares DKIM's ease of deployment: they both use the DNS to distribute a single public key for each domain. Importantly, LES supports common uses of email that DKIM jeopardizes: multiple email personalities, firewalled ISPs, incoming-only email forwarding services, and other common uses that often require sending email via a third-party SMTP server. In addition, LES does not require DKIM's implied intra-domain mechanism for authenticating users when they send email. LES provides these features using identity-based signatures. Each domain authority generates a master keypair, publishes the public component in the DNS, and stores the private component securely. Using this private component, the authority delivers to each of its users, via email, an individual secret key whose identity string corresponds to the user's email address. A sender then signs messages using this individual secret key. A recipient verifies such a signature by querying the appropriate master public key from the DNS, computing the sender's public key, and verifying the signature accordingly. As an added bonus, the widespread availability of user-level public keys enables deniable authentication, such as ring signatures. Thus, LES provides email authentication with optional repudiability. We built a LES prototype to determine its practicality. Basic user tests show that the system is relatively easy to use, and that cryptographic performance, even when using deniable authentication, is well within acceptable range.