TAMU
rtds

Real Time Distributed Systems Lab
Research
Members
Publications
Contact
Home
Research
Systems & Architecture
Bio & Medical

Phantom Bulk Email Generator (PBEG)

A major limiting factor of existing Unsolicited Bulk Emails (UNBE) filters is their inability in defeating the rapidly evolved designs. To support objective evaluation of UNBE detection algorithms, we are developing an UNBE workload generator to produce phantom UNBE emails. The system uses a database to manage generation of UNBE features , e.g., URL links that can be inserted into the phantom emails.

Main capabilities

  • UNBE features: Traces collected from spamming email repositories can be used to extract instances of UNBE features, e.g., URL links. These instances can be used as the seeds in (phantom) spamming email generation.
  • UNBE generation: Creation of UNBE copies based on user specified parameters (structures of the UNBE copies, their arrival sequence, workload sizes, etc.)
  • Experiment control: Mixture of UNBE and regular emails based on specific distributions.
  • XML based experiment scripts : The experiment management software wraps input parameters in the XML format and sends it to the generator.

Bulk EMail Generation System

Figure 1. BMG System Architecture.

The three major subsystems in the BMG consist of (1) the MIME parser to extract the UNBE features, such as URL links, (2) the mutator to compose email instances, and (3) the MIME generator to create a new MIME email. The new email is then sent to the UNBE storage for future detection. The UNBE samples generated by our system will be made available to the research community in the near future.
UNBE download coming soon.