Spectrum-aware data replication in intermittently connected cognitive radio networks

The opening of under-utilized spectrum creates an opportunity for unlicensed users to achieve substantial performance improvement through cognitive radio techniques. In cognitive radio ad-hoc networks, with node mobility and low node density, the network topology is highly dynamic and end-to-end connection is hard to maintain. We propose data replication techniques to address these problems and improve data access performance in such intermittently connected cognitive radio network. Although data replication has been extensively studied in traditional disruption tolerant networks, existing techniques cannot be directly applied here since they do not consider the effects of primary user appearance on data replication. In this paper, we formulate spectrum-aware data replication as an optimization problem which tries to maximize the average data retrieval probability, subject to storage and time constraints. Since the problem is hard to solve based on mixed integer programming, we further design a distributed replication scheme based on the metric of replication benefit. Extensive simulations based on synthetic and realistic traces show that our scheme outperforms existing schemes in terms of data retrieval probability in various scenarios.