Abstract :
Recently, the use of a virtual backbone (connected dominating set) in various applications in mobile ad hoc networks (MANETs) has become popular. These applications include topology management, point and area coverage, and routing protocol design. In a mobile environment such as a MANET, one challenging issue in constructing a virtual backbone is to accomplish a distributed and localized solution that aims at balancing several contradicting objectives: small approximation ratio, fast convergence, and low computation cost. Many existing distributed and localized algorithms select a virtual backbone without resorting to global or geographical information. However, these algorithms incur a high computation cost in a dense network. In this paper, we propose a distributed solution based on reducing the density of the network using two mechanisms: clustering and adjustable transmission range. By using adjustable transmission range, we also achieve another objective, energy-efficient design, as a by-product. As an application, we show an efficient broadcasting where nodes (and only nodes) in a virtual backbone are used to forward the broadcast message. The virtual backbone is constructed using Wu and Li´s marking process [J. Wu et al., (1999)] and the proposed density reduction process. The application of the density reduction process to other localized algorithms is also discussed. The efficiency of our approach is confirmed through both analytic and simulation study.
Keywords :
ad hoc networks; computational complexity; distributed algorithms; message passing; mobile computing; routing protocols; workstation clusters; MANET; adjustable transmission range; broadcast message; connected dominating set; distributed algorithm; mobile ad hoc networks; routing protocol design; virtual backbone; Broadcasting; Clustering algorithms; Computational efficiency; Computer networks; Distributed computing; Mobile ad hoc networks; Mobile computing; Network topology; Routing protocols; Spine;