TRUMP: Supporting efficient realization of protocols for cognitive radio networks

Cognitive radios require fast reconfiguration of the protocol stack for dynamic spectrum access and run-time performance optimization. In order to provide rapid on-the-fly adaptability of PHY/MAC protocols, we have designed and implemented TRUMP: a Toolchain for RUn-tiMe Protocol realization. It includes a meta-language compiler, logic controller and an optimizer. TRUMP allows run-time realization and optimization of cognitive network protocols for the requirements of a particular application, communication capabilities of the radio, the current spectrum regulation and policies. TRUMP supports efficient multi-threading for multi-core platforms in order to meet variable computational requirements and to allow parallelization of PHY/MAC processing for cognitive radio systems. We have carried out the performance evaluation for different metrics on WARP SDR platform and embedded Linux based PCs. Our results indicate that TRUMP allows reconfiguration of protocols in the order of a few microseconds through run-time linking of different components, thus meeting the strict timeliness requirements imposed by PHY/MAC processing.