A Spatio-Temporal Approach to Selective Data Dissemination in Mobile Peer-to-Peer Networks

We examine data dissemination in mobile peer-to-peer networks, where moving objects communicate with each other via short-range wireless technologies such as IEEE 802.11 or Bluetooth. Given the memory and bandwidth/energy constraints at moving objects, the ideal mobile peer-to-peer dissemination method is for each moving object to store and transmit only the reports that are new to other objects encountered in the future. However, in practice no object can know the states of all the other objects due to the distributed and dynamic nature of mobile peer-to-peer networks. Thus, predicting the novelty probability of a report is important for efficient data dissemination in mobile peer-to-peer networks. In this paper we propose a decentralized spatio-temporal approach to selective data dissemination. In this approach, the novelty probability of a report is estimated based on both spatial and temporal attributes (AGE and DISTANCE) of the report and each moving object only stores and transmits the reports with the highest novelty probabilities. We study different strategies of using novelty factors to compute the novelty probability. Extensive experiments are conducted to test and analyze the performance of different strategies. The experimental results determine the best strategy and demonstrate its superiority against existing mobile peer-to-peer methods.