Abstract

Cognitive Radio (CR) technology is considered as a promising technology to overcome
spectrum scarcity problem in wireless networks, by sharing the spectrum between both unlicensed
users (secondary users, (SUs)) and licensed users (primary users, (PUs)), provided that the SUs
respect the PUs’ rights to use the spectrum exclusively.
An important technical area in cognitive radio networks (CRNs) is wireless security. A
secure CRN must meet different security requirements, which are: confidentiality, integrity,
availability and authentication. Data confidentiality is a mandatory requirement in cognitive radio
networks, generally to maintain the privacy of the data owner (PU or SU). Integrity means that
data is transmitted from the source to the destination without alteration. While availability is to
release the channels assigned to one SU as soon as a PU wants to use its spectrum. Authentication
in CRN means that each node has to authenticate itself before it can use the available spectrum
channels.
New classes of security threats and challenges in CRNs have been introduced that target
the different layers of OSI model and affect the security requirements. Providing strong security
may prove to be the most difficult aspect of making CR a long-term commercially-viable concept.
Protection of routes used for data transmission is a critical prerequisite to ensure the robustness of
iv
the routing process. Therefore, route discovery must be done in such a way that lets each node find
the best secure path(s) for its data transmission.
In this work, network security of CRN is improved through proposing different models that
are built to fulfil the security requirements mentioned above. Improving the network security
enhances the network performance, taking into consideration the quality of service (QoS) desired
by the different network nodes such as bandwidth and time delay. This work aims to combine the
spectrum sensing phase and the spectrum management phase, as well as to detect all the adversary
nodes that slow down the network performance by selectively holding and not forwarding packets
to their next hop(s). We measure the network node’s reliability for using network resources
through a value called belief level (BL), which is considered as the main parameter for our entire
work. BL is used to monitor the nodes’ behavior during the spectrum sensing phase, and then it is
used to form the best path(s) during the spectrum management phase. Particularly, this work
follows a hierarchical structure that has three different layers. At the bottom layer, a novel
authentication mechanism is developed to fulfil the authentication and the availability security
requirements, which ends assigning a belief level (BL) to each node. At the middle layer, the
nodes’ behavior during the spectrum sensing phase is monitored to detect all the adversary node(s).
Finally, at the top layer, a novel routing algorithm is proposed that uses the nodes’ security (BL)
as a routing metric. SUs collaborate with each other to monitor other nodes’ behavior. Users’ data
confidentiality and integrity are satisfied through this hierarchical structure that uses the cluster-based,
central authority, and nodes collaboration concepts. By doing so, the traffic carried in the
CRN is secured and adversary nodes are detected and penalized.