On the Trade-offs of Cross-Layer Protocols for Cognitive Radio Networks
Résumé
The design of future wireless networks is currently investigating the benefits of cognition in terms of quality of service and adaptability to the current working conditions. In such cognitive networks, intelligent services are implemented in the wireless devices which enables them to maximize their own capabilities and ultimately, yield an improvement in the global network performance. These services mostly include cooperative decision mechanisms at the lower layers of the protocols stack where wireless nodes adjust their own transmission parameters (e.g. transmission power or channel) knowing their neighbors' settings. To increase the benefits of such cooperative behavior in terms of global networking performance, cross-layer design can be implemented to adjust the decisions of higher layer protocols according to the lower layer modifications. Our perspective is to show how and when such cross-layer design can improve the overall performance of cognitive networks and what are the trade-offs in terms of complexity versus energy gains that arise in this case. For this purpose, we present a typical cross-layer iterative optimization of the routing, medium access and physical layers. In this implementation, the price to pay for cross-layering mostly resides in the overhead due to the iterative optimization of routing and lower layers' parameters.