A Leader Election Protocol for Cognitive Radio Networks

Leader election is a fundamental problem of distributed computing systems. In cognitive radio network (CRN), the secondary users (SUs) are connected under the leased spectrum of primary user (also called, licensed user) and hence often called opportunistic network. The emerging trend is to maximize the channel utilization in CRN. However, the computational activities performed by the SUs depend on the activity of primary user. Thus, in general, CRN is highly dynamic and network architectures are short lived. Many applications require a leader node to carry out better coordination among the participating nodes. The CRN being a highly dynamic network, the leader election is more challenging than in other networks. The leader node coordinates the activities of SUs and regulates the appropriate channel among them keeping in view the behavioral activities of PUs, which leads to enhanced channel utilization. We propose a diffusion computation based leader election protocol for CRN. The protocol is "weakly" self stabilizing and terminating.