Highly Scalable and Efficient Publish/Subscribe Protocols Using Geographic Information for Wireless Sensor Networks

Publish/subscribe (Pub/Sub) systems have been used in traditional distributed computing applications to provide effective and efficient event services. Recently, the design of Pub/Sub systems are proposed for wireless sensor networks (WSNs). Some of the proposed Pub/Sub protocols address the issue of establishing event delivery path with low routing cost by using geographic information. However, they have not considered how to reduce redundant event delivery, which may cause high resource consumption and poor system scalability. In this paper, we propose protocols which address both issues. More specifically, our protocols use geographic information to save routing cost, and meanwhile reduce redundant event delivery by letting the subscribers share the event delivery paths. Three highly scalable and efficient Pub/Sub protocols for WSNs, namely shortest delivery path (SDP), shortest delivery path with lower delivery overhead (SDP-LDO), and lowest delivery overhead (LDO), are designed to achieve different performance goals. The results of theoretical and experimental evaluation show that the proposed protocols can significantly improve the resource efficiency and scalability of a Pub/Sub system compared to the previous solutions. Our protocols can also be used to achieve a good tradeoff between the costs of event subscribing and event delivery.