FlexiCast: Fast Data Sharing in Skewed Wireless Sensor Networks

Abstract

Efficient sharing of data is a key component in any decentralized system. There have been quite sig- nificant developments in the data-sharing protocols, especially with the advent of the synchronous/concurrent transmission based strategies. However, most of these existing data sharing strategies are tested either in small scale simulated networks or in testbed settings with fixed usual homogeneous structures. Real-life IoT/WSN networks, in contrast, often bear diverse structural patterns due to a highly skewed distribution of the nodes, e.g., highly dense in some places, while sparsely distributed along pathways in some other places. Our study reveals that that while the existing state-of-the-art protocols perform very well in uniformly distributed networks, they dwindle in face of these skewed networks commonly found in real life. Lack of the ability of these protocols to self-adjust as per the underlying network structure is one of the prime reasons for such performance degradation. In this work, we design and implement a flexible and self-adjusting many-to-many communication protocol FlexiCast and demonstrate that it consistently performs better than the current state-of-the-art protocols in a wide variety of networks. We show that FlexiCast can achieve all-to-all data sharing up to 1.77 times faster while consuming up to 1.87 times lesser energy compared to the state-of-the-art protocols.