A Data Mining Based Publish/Subscribe System over Structured Peer-to-Peer Networks - Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing - page 1

A Data Mining Based Publish/Subscribe System
over Structured Peer-to-Peer Networks

Junping Song, Haibo Wang, Pin Lv, Shangzhou Li, and Menglu Xu

Abstract. In this paper, we propose a data mining based publish/subscribe system
(DMPSS). First, the data mining technology is used to find attributes that are usu-
ally subscribed together, e.g. frequent itemset. Then subscriptions and events are
installed by frequent itemsets contained in them. If subscriptions and events don't
contain any frequent itemset, they are delivered to specified RPs (rendezvous points)
for matching. The usage of frequent itemsets provides two advantages to DMPSS.
First, it achieves even matching load distribution on RPs. Second, it reduces the
event publication cost. The performance of DMPSS is evaluated by simulations.
The experimental results show that DMPSS realizes even matching load distribu-
tion, and it reduces the overhead for message transmission and latency dramatically.

1 Introduction

The publish/subscribe system is widely used for delivering data from publishers
(data producers) to subscribers (data consumers) across large-scale distributed net-
works in a decoupled fashion. Traditional publish/subscribe applications include
news and sports ticker services, real-time stock quotes and updates, market tracker,
etc. [1] In recent years, the publish/subscribe scheme has been used in many Web
2.0 applications, especially On-line Social Networks like Twitter, Facebook and
Google+ [2]. However, most of the above mentioned applications adopt topic-based
subscriptions, which offer very limited expressiveness to subscribers. Currently, the
study of content-based systems, which allow fine-grained subscriptions by enabling
restrictions on the event content, has attracted plenty of attention [3-7].