Full Reversal Routing as a Linear Dynamical System
Link reversal is a versatile algorithm design paradigm, originally proposed by Gafni and Bertsekas in 1981 for routing, and subsequently applied to other problems including mutual exclusion and resource allocation. Although these algorithms are well-known, until now there have been only preliminary results on time complexity, even for the simplest link reversal scheme for routing, called Full Reversal (FR). In this paper we tackle this open question for arbitrary communication graphs. Our central technical insight is to describe the behavior of FR as a dynamical system, and to observe that this system is linear in the min-plus algebra. From this characterization, we derive the first exact formula for the time complexity: Given any node in any (acyclic) graph, we present an exact formula for the time complexity of that node, in terms of some simple properties of the graph. These results for FR are instrumental in analyzing a broader class of link reversal routing algorithms, as we show in a companion paper that such algorithms can be reduced to FR. In the current paper, we further demonstrate the utility of our formulas by using them to show the previously unknown fact that FR is time-efficient when executed on trees.
Unable to display preview. Download preview PDF.
- 6.Attiya, H., Gramoli, V., Milani, A.: A provably starvation-free distributed directory protocol. In: 12th International Symposium on Stabilization, Safety, and Security of Distributed Systems, pp. 405–419 (2010)Google Scholar
- 7.Park, V.D., Corson, M.S.: A highly adaptive distributed routing algorithm for mobile wireless networks. In: 16th Conference on Computer Communications (Infocom), apr 1997, pp. 1405–1413 (1997)Google Scholar
- 8.Ko, Y.-B., Vaidya, N.H.: Geotora: a protocol for geocasting in mobile ad hoc networks. In: Proceedings of the 2000 International Conference on Network Protocols, ICNP 2000, pp. 240–250 (2000)Google Scholar
- 12.L., J., Malpani, N.V.N., Welch: Leader election algorithms for mobile ad hoc networks. In: Proceedings of the 4th International Workshop on Discrete Algorithms and Methods for Mobile Computing and Communication (2000)Google Scholar
- 14.Ingram, R., Shields, P., Walter, J.E., Welch, J.L.: An asynchronous leader election algorithm for dynamic networks. In: Proceedings of the IEEE International Parallel & Distributed Processing Symposium, pp. 1–12 (2009)Google Scholar
- 15.Busch, C., Surapaneni, S., Tirthapura, S.: Analysis of link reversal routing algorithms for mobile ad hoc networks. In: Proceedings of the 15th ACM Symposium on Parallelism in Algorithms and Architectures (SPAA), pp. 210–219 (2003)Google Scholar
- 17.Charron-Bost, B., Gaillard, A., Welch, J.L., Widder, J.: Routing without ordering. In: Proceedings of the 21st ACM Symposium on Parallelism in Algorithms and Architectures (SPAA), pp. 145–153 (2009)Google Scholar
- 19.Charron-Bost, B., Függer, M., Welch, J.L., Widder, J.: Partial is full. In: Kosowski, A., Yamashita, M. (eds.) SIROCCO 2011. LNCS, vol. 6796, pp. 111–123. Springer, Heidelberg (2011)Google Scholar
- 20.Charron-Bost, B., Függer, M., Welch, J.L., Widder, J.: Full reversal routing as a linear dynamical system. Research Report 7/2011, Technische Universität Wien, Institut für Technische Informatik, Treitlstr. 1-3/182-2, 1040 Vienna, Austria (2011)Google Scholar
- 21.Heidergott, B., Olsder, G.J., von der Woude, J.: Max plus at work. Princeton Univ. Press, Princeton (2006)Google Scholar