Concept Tree-Based Event Matching Algorithm in Publish/Subscribe Systems

Abstract

Semantic-based publish/subscribe system has attracted a lot of attention in recent years due to its powerful description ability in message dissemination scenarios. As a key part of semantic-based publish/subscribe systems, event matching needs to understand the semantic meaning of subscriptions, especially the hierarchy of concepts. However, existing event matching algorithms are severely affected by the complexity of concept hierarchy trees, some even cannot run due to high memory occupation. This article proposes an event matching algorithm called CTPS (concept tree-based publish/subscribe system) to address it; specifically, we build subscription indexes on concept hierarchy trees for the first time, employ bit arrays to avoid unnecessary matches, and use faster bit operations to accelerate the matching speed. Experiments show that, compared with previous algorithms, CTPS has less memory occupation and shorter event matching time, and its performance is not sensitive to the height of the concept hierarchy tree.