Preventing Spam in Publish/Subscribe

Abstract

Publish/subscribe protocols are becoming popular on the Internet with the advent of RSS-feeds and Session Initiation Protocol (SIP) events. Since publish/subscribe is by nature many-to-many form of communication, it is vital to prevent spam and bogus messages. This problem has not yet been fully addressed. In this paper, we investigate the spam problem in distributed content-based publish/ subscribe overlay networks and outline several solutions inspired by current email spam prevention techniques. We analyze filter-based spam in distributed pub/sub systems and focus on the detection and mitigation of bogus publishers and brokers.