Skip to Main Content
This paper analyzes a spam rejection scheme at Simple Mail Transfer Protocol (SMTP) sessions. This scheme utilizes a layer-3 e-mail pre-classification technique to estimate e-mail classes before an SMTP session ends. We study the spam rejection scheme using discrete-time Markov chain analysis and analyze the performance of the proposed scheme under different e-mail traffic loads and service capacities. The proposed scheme reduces the e-mail volume to be queued and processed by e-mail servers. This reduces non-spam e-mail queuing delay and loss, and protects e-mail servers from being overloaded by spam traffic.