Preventing Outgoing Spam by Monitoring User Interactions

Abstract

Filtering unsolicited messages in recipients' inboxes and restricting them at the mail server or transit agent level are two practical ways to confine spam. However, none of them prevents spam messages from leaving the source machines and flooding the Internet. In this paper, we present a method of preventing outgoing spam from entering the network by correlating user activities with the outbound network traffic. The message intentability scoring algorithm (MISA) is proposed to assess the intentability of the outgoing messages and an outgoing spam prevention system is designed based on this algorithm. This idea can be used to build standalone outgoing spam prevention or an extension of host-based intrusion prevention system