A Performance Evaluation of Flooding Methods on Real Network Topologies

In order to ensure the efficient distribution of information, routing protocols, data collection and dissemination mechanisms are essential. Flooding is a commonly used method to broadcast information to all nodes in a network, with various algorithms having different advantages and disadvantages. In this paper, an evaluation of three flooding methods is presented: (i) blind flooding; (ii) probabilistic flooding; and (iii) m-Probabilistic flooding. The analysis focuses on the achieved coverage and the number of messages, showing the performance of each flooding method on these aspects. Findings in this paper reveal that, in certain cases, m-Probabilistic flooding achieves better coverage than the other considered methods and significantly reduces the number of exchanged messages. Along with the comparison results, a metric that is based on binomial approximation with respect to topologies is also investigated to shed further light on the particulars of probabilistic flooding.