Subscription email

David D Lowe


Supervised by Frank C Langbein; Moderated by Ralph Martin

Project aims and objectives: To explore subscription-only Internet messaging protocols like IM2000, to design a new protocol and to implement a proof-of-concept prototype of a server and mail user agent.

Description: Email in its current implementations suffers from many shortcomings. Its two most glaring problems are its inherit insecurity and the huge volume of spam.

Many solutions have been proposed to tackle these problems, but all of them fall short. The risk of false positives in spam filters is still too high, and the cost of spamming still too low. SSL doesn't necessarily encrypt messages from end-to-end, and PGP is too cumbersome for rapid adoption by end-users. A completely different architecture is needed.

IM2000 and similar mail architectures specify that messages be stored on the sender's mail server rather than on the recipient's. Every user would maintain a contact list that would serve as list of servers to poll for new messages. This would eliminate most spam, as one would only be able to receive messages from pre-approved addresses. Every user would automatically be assigned encryption key pairs, and the implementation would allow for automatic distribution of public keys, solving the usability problems behind traditional PGP security.

The implications of this new architecture would be explored, and the opportunity to fix some of SMTP's other design flaws would be considered.

Resource requirements: At least two servers would be needed to implement the prototype. These servers could be virtual.

Initial Plan (22/10/2012) [Zip Archive]

Interim Report (14/12/2012) [Zip Archive]

Final Report (03/05/2013) [Zip Archive]

Publication Form