Please use this identifier to cite or link to this item:
|Title:||Analysis of a high-performance mac protocol for underwater acoustic networks||Authors:||Shahabudeen, S.
|Keywords:||Ad hoc underwater networks
medium-access control (MAC)
multiple access with collision avoidance (MACA)
service time distribution
|Issue Date:||Jan-2014||Citation:||Shahabudeen, S., Motani, M., Chitre, M. (2014-01). Analysis of a high-performance mac protocol for underwater acoustic networks. IEEE Journal of Oceanic Engineering 39 (1) : 74-89. ScholarBank@NUS Repository. https://doi.org/10.1109/JOE.2013.2246741||Abstract:||Multiple access with collision avoidance (MACA) is a popular medium-access control (MAC) protocol for terrestrial networks (e.g., 802.11). Underwater acoustic networks (UANs) differ from terrestrial networks as they are characterized by long propagation delays and higher data loss. Most of the published analysis for MACA in terrestrial networks ignores these factors, and, therefore, cannot be directly applied to UANs. As a result of the high data loss in UANs, it is common to implement reliability at a link level, rather than end-to-end acknowledgements. Keeping this in mind, we present a Markov chain analysis for a reliable variant of the MACA protocol for ad hoc UANs and derive closed-form expressions for mean service time and throughput. We show that the network performance is vastly improved with a few changes to the protocol, and propose a novel MACA-based protocol for use in UANs. For best performance, the protocol parameters such as batch size and backoff window have to be optimally chosen. We show that an optimum batch size minimizes the total waiting time. Finally, we compare our analysis results with experimental results obtained from a deployment of a small UAN. © 2013 IEEE.||Source Title:||IEEE Journal of Oceanic Engineering||URI:||http://scholarbank.nus.edu.sg/handle/10635/81960||ISSN:||03649059||DOI:||10.1109/JOE.2013.2246741|
|Appears in Collections:||Staff Publications|
Show full item record
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.