Scalable and robust data dissemination for large-scale wireless sensor networks

摘要

The publish/subscribe paradigm has been known as the appropriate communication model for providing high scalability and mobility in data-centric networks such as wireless sensor networks. Through an internal infrastructure, in the publish/subscribe paradigm consumers and producers are completely decoupled in the three essential properties: decoupling in space, time, and synchronization. In other words, the internal infrastructure should be hidden from consumers and producers, and it should provide anonymous, asynchronous, and multicasting communication for the decoupling properties. The properties are suitable for the large-scaled mobile environment with many mobile consumers and many mobile producers. Hence, we propose a novel data dissemination scheme satisfied fully with three decoupling properties, and an embedded grid structure is used as the internal infrastructure for supporting shared buffering and multicasting. Through our performance analysis from computer simulation, we conclude that our scheme may be not only the first attempt to implement the true publish/subscribe paradigm in wireless sensor networks, but also can support high scalability and mobility in the wireless sensor networks.