A Simulation-based Performance Comparison of the Minimum Node Size and Stability-based Connected Dominating Sets for Mobile Ad Hoc Networks

The high-level contribution of this paper is a simulation-based comparison of two contrasting categories (minimum node size vs. stability) of connected dominating sets for mobile ad hoc networks. We pick the maximum density-based CDS (MaxD-CDS) and Node ID-based CDS (ID-CDS) to be representatives of the minimum node size-based CDS algorithms; the minimum velocity-based CDS (MinV-CDS) and the node stability index-based CDS (NSI-CDS) are chosen as representatives for stability-driven CDS. The MaxD-CDS, ID-CDS and MinV-CDS algorithms prefer to respectively include nodes with a larger number of uncovered neighbors, larger node ID and lower velocity into the CDS; the NSI-CDS algorithm prefers to include nodes with a larger value for the sum of the predicted expiration times of the links with the neighbor nodes. We simulate these algorithms under diverse conditions of node mobility and network density. We observe the MinV-CDS to be the most stable among all the CDSs as well as incur a lower hop count per path (through the CDS nodes) between any two nodes in the network. However, the MinV-CDS incurs a relatively larger CDS node size and edge size compared to the other three CDSs. Owing to a larger control overhead that could be incurred while broadcasting through a MinV-CDS (with larger number of CDS nodes and edges), the NSI-CDS can be considered as the best choice from the points of view of delay, energy, bandwidth and fairness of node usage.