p2p ToPSS

Peer-to-Peer-Based Toronto Publish/Subscribe System

University of Toronto

ECE Department

Overview

Building distributed content-based publish/subscribe systems has remained a challenge. Existing solutions typically use a relatively small set of trusted computers as brokers, which may lead to scalability concerns for large Internet-scale workloads. Moreover, since each broker maintains state for a large number of users, it may be difficult to tolerate faults at each broker. In this project we propose an approach to building content-based publish/subscribe systems on top of distributed hash table (DHT) systems. DHT systems have been effectively used for scalable and fault-tolerant resource lookup in large peer-to-peer networks. Our approach provides predicate-based query semantics and supports constrained range queries. Experimental evaluation shows that our approach is scalable to thousands of brokers, although proper tuning is required.

DHT Substrate

We implemented our approach on top of Pastry. Our choice of platform was not fundamental to our design and we could have built our system on top of any DHT system.

People

Faculty: Graduate Students:

Publications

David Tam, Reza Azimi, Hans-Arno Jacobsen, "Building Content-Based Publish/Subscribe Systems with Distributed Hash Tables", International Workshop on Databases, Information Systems and Peer-to-Peer Computing, September 7-8, 2003, Humboldt University, Berlin, Germany. Here is the official Springer-Verlag hyperlink [ MetaPress Direct Linking | OpenURL ]

Slides

Workshop Presentation [pdf | ppt]: International Workshop on Databases, Information Systems and Peer-to-Peer Computing, September 7-8, 2003, Humboldt University, Berlin, Germany.

Source Code

We implemented our ideas using FreePastry.
David Tam
Last modified: Fri Aug 19 16:15:26 EDT 2005