A Peer-to-Peer Architecture for Collaborative Spam Filtering

Abstract

Over the last twenty years, the importance of electronic mail has risen to a point where it is now a critical medium for many companies and individuals. This fact has led to the emergence of a new method of invasive marketing: Unsolicited Commercial Email, otherwise known as spam. Spam is costing email users an increasing amount of time and money.

It is the act of blindly mass mailing a message that makes it spam, and not merely its content, making it harder to spot automatically. Because of this, the need for a flexible, precise, cheap and easy to use spam filter has yet to be met. A recent idea that has proved promising for accurate detection of spam is the collaborative sharing of knowledge about spam between users.

This thesis describes a distributed peer to peer architecture that allows users to store and share data about spam. This design provides a trust management model that offers both precise filtering mechanisms and protection from malevolent users. One possible implementation of this architecture is also proposed, and used for deterministic evaluation of the design.