SpamWatch icon news | downloads | documentation | links 

SpamWatch - A Peer-to-peer Spam Filtering System

SpamWatch is a collaborative spam filtering system built on top of Tapestry, a peer-to-peer location/routing infrastructure. SpamWatch has the following features:
  • Collaborative - Every user in a SpamWatch community contributes to the community by tagging certain emails as spam, and benefits from the contribution of all others.
  • Peer-to-peer - Peer nodes (often local mail servers) join the large peer-to-peer network and provide service to local users. Tapestry ensures quick and accurate location of records and supports dynamic joining/leaving of peers.
  • Content similarity based - The identification of spam emails are based on the content of emails, not on subject, sender or IP address. Moreover, moderate altercation of the content is tolerated. Common techniques used by spammers such as changing the salutation or spacing between words will not prevent the spams from being caught.
Source code and binary files are available below for download. For more details about how SpamWatch works, you can read our ACM Middleware 2003 paper available below.

Note to end users: SpamWatch should be considered alpha-quality software because of its research prototype nature. There's currently no long-running SpamWatch servers. Refer to README files of the server and client for instructions of setting your own server up and how to use it.

SpamWatch is a Berkeley CS262a course project by Feng Zhou and Li Zhuang, with great help from Ben Zhao and Ling Huang.


Mar 27, 2004 SpamWatch is mentioned by Popular Science!
Mar 31, 2003 ACM Middleware 2003 paper posted.
Jan 4, 2003 Overview poster and Outlook addin screenshot posted.
Dec 13, 2002 SpamWatch Server (WatchTower) and Outlook Plugin 0.1 source code and binaries posted.
Dec 2, 2002 Web page up.



Client: SpamWatch Outlook Plugin 0.1 For Microsoft Outlook 2000/XP README
Server: WatchTower 0.1 (SpamWatch Server on Tapestry) For Linux(glibc 2.1) & IBM JDK 1.3 README




  • CS262A Fall 2002 - Advanced Topics in Computer Systems
  • Tapestry - Infrastructure for Fault-resilient, Decentralized Location and Routing
  • Oceanstore - Providing Global-Scale Persistent Data
Maintainer: Feng Zhou, 2002-2003 Last updated: $Date: 2004/03/28 00:10:21 $