Distributed Construction of D-Hop Connected Dominating Sets for Wireless Sensor Networks

Several critical operations such as, data collection, routing, service discovery, etc., employ various types of information dissemination in order to be carried out. To avoid costly flooding-based solutions reaching out to all nodes, it is frequently sufficient that only a proper subset of the nodes (or backbone network) be involved, i.e., the one ensuring that every other node will be at most d hops away from a node belonging to the said subset of nodes. Finding this subset is equivalent to the construction of a d-hop Connected Dominating Set (d-CDS). Given the high complexity (non-polynomial) and the requirement of global information for constructing a Minimum d-CDS (d-MCDS), in this paper a distributed algorithm that relies on local information (i.e., d hops away) to construct an approximation of the d-MCDS is developed. The proposed algorithm is studied and compared against a centralized one that is an approximate solution for the minimum d-MCDS problem, and a recently proposed distributed one. It is shown that the size of the constructed d-CDS under the proposed algorithm is (i) close and sometimes smaller than that under the recently proposed one; and (ii) close to the centralized one. In addition, the number of transmitted messages is significantly reduced under the proposed algorithm, which is important for preserving energy resources in wireless sensor networks.