Data-aware multicast

Abstract

This paper presents a multicast algorithm for peer-to-peer dissemination of events in a distributed topic-based publish-subscribe system, where processes publish events of certain topics, organized in a hierarchy, and expect events of topics they subscribed to. Our algorithm is "data-aware" in the sense that it exploits information about process subscriptions and topic inclusion relationships to build dynamic groups of processes and efficiently manage the flow of information within and between these process groups. This "data-awareness" helps limit the membership information that each process needs to maintain and preserves processes from receiving messages related to topics they have not subscribed to. It also provides the application with means to control, for each topic in a hierarchy, the trade-off between the message complexity and the reliability of event dissemination. We convey this trade-off through both analysis and simulation.