ROSTER: Radio Context Attestation in Cognitive Radio Network.

In cognitive radio network, cognitive radios dynamically reconfigure themselves based on the spectrum opportunity. It is envisioned to be the key of overcoming the spectrum shortage. However, such reconfigurability also amplifies potential harmful interferences from non-compliant radios. In this paper, we propose ROSTER, a radio context attestation protocol for cognitive radio network. The proposed protocol is based on our observation that the compliance of a radio transmission depends on software configuration, radio configuration as well as the location and time of the device, which we call radio context. We believe radio context attestation, which allows the authority to verify the operational integrity of individual cognitive radio, is a fundamental security function for cognitive radio networks. To the best of our knowledge, we are the first to study this important problem. Different from conventional software attestation, ROSTER is designed to handle dynamic configurations in cognitive radios. Furthermore, special considerations are given in the protocol design to accommodate different levels of sensitivity in spectrum databases. Besides protocol design and security analysis, we also build a prototype of the proposed system using Raspberry Pi, USRP, and Amazon AWS. Network simulation using the benchmark measurements from the prototype shows the scalability of our proposed protocol.