A Measurement Study of a Publish Subscribe System

While publish-subscribe systems have attracted much research interest in the last decade, few established benchmarks have emerged and there has been little characterization of how they are used in practice. This paper examines RSS, a newly emerging, widely used publishsubscribe system for web micronews. Based on a trace study spanning 45 days at a medium-size academic department, and periodic polling of approximately 100,000 RSS feeds, we extract feed and client characteristics for RSS. We find that the popularity of RSS feeds follows a power law distribution. 16% of RSS feeds are updated hourly and 25% do not change at all during our polling period. 64% of all updates involve less than three lines in the XML document. We also find that RSS feed update sizes are proportional to feed size. Overall, an analysis of RSS, the first widely deployed publish-subscribe system, can help inform the design of next generation pub-sub systems.