Dynamic event subscriptions in distributed event based architectures

Abstract

The burst of wireless sensors networks transmitting events paved the way for infrastructures that link uniquely identifiable things to their virtual representations in the Internet. As the Internet of Things gains momentum for leveraging and offering information access to service-based applications and users, challenges associated with efficiently publishing, subscribing, processing and reacting to events emerge. This work focuses on the challenge of enabling efficient active capability in large, distributed event infrastructures. We aim to support dynamic recommendations of new event streams to which a service should subscribe for a meaningful period of time, in order to take advantage of situational information. We present a goal-driven, ECA-based hierarchical model, called Situation-Action-Network (SAN) and we implement the Event Subscription Recommender (ESR) software that uses SANs in order to produce dynamically new event subscriptions or cancel existing ones before they become obsolete, based on detected situations. We evaluated ESR in a scenario that involves events transmitted by marine vessels in order to efficiently enforce safety regulations on behalf of port authorities. The results of this evaluation were compared against a traditional ECA-based system and we found that ESR outperformed existing systems with respect to subscription efficiency.