ZePoP: A Distributed Leader Election Protocol using the Delay-based Closeness Centrality for Peer-to-Peer Applications

Abstract

This paper presents ZePoP, a leader election protocol for distributed systems, optimizing a delay-based closeness centrality. We design the protocol specifically for the Peer to Peer(P2P) applications, where the leader peer (node) is responsible for collecting, processing, and redistributing data or control signals satisfying some timing constraints. The protocol elects an optimal leader node in the dynamically changing network and constructs a Data Collection and Distribution Tree (DCDT) rooted at the leader node. The elected optimal leader is closest to all nodes in the system compared to other nodes. We validate the proposed protocol through theoretical proofs as well as experimental results.