Self-organizing Mobile Ad Hoc Networks: Spontaneous Clustering at the MAC Layer J. Alonso-Z´ arate 1 , E. Kartsakli 2 , P. Chatzimisios 3 , L. Alonso 2 , and Ch. Verikoukis 1 1 Centre Tecnol` ogic de Telecomunicacions de Catalunya (CTTC) Av. del Canal Ol´ ımpic s/n, CTTC, 08860, Castelldefels, Barcelona, Spain {jesus.alonso,cveri}@cttc.es 2 Dept. of Signal Theory and Communications, Universitat Polit` ecnica de Catalunya (EPSC-UPC) Av. del Canal Ol´ ımpic s/n, EPSC, 08860, Castelldefels, Barcelona, Spain {ellik,luisg}@tsc.upc.edu 3 Dept. of Technology Management, University of Macedonia GR-59200, Nousa, Greece pchatzim@uom.gr Abstract. We present in this paper a master-slave, self-organized, spon- taneous, passive, and dynamic clustering algorithm embedded into the Medium Access Control (MAC) layer for Mobile Ad hoc Networks. Any mobile station gets access to the channel by executing a contention-based mechanism similar to the IEEE 802.11 Standard. However, once it seizes the channel, it establishes a temporary cluster to which closer neigh- bors can get synchronized. Within each cluster, any infrastructure-based MAC protocol can be executed. Link-level computer simulations have been carried out to show that this approach can remarkably improve the performance of ad hoc networks at the MAC layer. Keywords: MAC, DQMAN, DQCA, Clustering, Ad hoc, Self-organizing. 1 Introduction Mobile ad hoc networks (MANETs) offer a set of advantages that look very promising to suit some of the hard and challenging requirements of a great number of new applications, which range from low-cost commercial systems or in- home applications, to rescue operations [1]. The need for continued connectivity and spontaneous networking has turned mobile ad hoc networking into a hot research topic over the last years. MANETs consist of a set of mobile stations that can freely move and want to communicate with each other. Since there is no central point of coordination, all the decisions should be made in a distributed manner and communications must be done by establishing peer-to-peer links between pairs of source and destination stations. In some cases, the source of information and its intended destination might not be within the same transmission range and, in this case, F. Granelli et al. (Eds.): MOBILIGHT 2009, LNICST 13, pp. 242–253, 2009. c  ICST Institute for Computer Sciences, Social-Informatics and Telecommunication Engineering 2009