Transparent cross-layer solutions for throughput boost in Cognitive Radio Networks

Cognitive Radio Network (CRN) is an emerging paradigm that makes use of Dynamic Spectrum Access (DSA) to communicate opportunistically, in the un-licensed Industrial, Scientific and Medical bands or frequency bands otherwise licensed to incumbent users such as TV broadcast. The opportunistic mode of data transfer introduces network-wide quiet periods for spectrum sensing and increased packet losses due to activities by Primary Users. TCP attributes packet delays and losses to congestion in the Internet, which does not perform well in wireless networks where the medium has inherent higher error rates than wired Internet. Opportunistic mode of communication in CRNs aggravates the already error-prone wireless communication resulting in further degradation of TCP performance due to higher latency and decreased throughput. This paper presents an analysis of TCP performance in IEEE 802.22 Wireless Regional Area Network (WRAN) based on Cognitive Radio Networks and proposes two approaches to improve the networking performance. The first approach makes base station resort to local recovery of lost frames between CRN base station and its clients, while the second approach implements a modified split TCP connection in which the base station sends crafted acknowledgements back to an Internet-side host on behalf of the corresponding CRN client to boost transmission speed. To the best of our knowledge, this is the first effort to study TCP performance in IEEE 802.22 based WRANs. Simulation results show that the proposed mechanisms result in improvement in TCP performance by as much as 20 times and conserve bandwidth by reducing retransmission overheads.