Logo eprints

Novel routing paradigms for wireless Mesh Networks

Nurchis, Maddalena (2011) Novel routing paradigms for wireless Mesh Networks. Advisor: Lenzini, Prof. Luciano. Coadvisor: Conti, Dr. Marco . pp. 156. [IMT PhD Thesis]

[img]
Preview
Text
Nurchis_phdthesis.pdf - Published Version
Available under License Creative Commons Attribution No Derivatives.

Download (1MB) | Preview

Abstract

The increasing desire of ubiquitous Internet access has recently promoted the deployment of wireless multi-hop networks in several application domains. Wireless Mesh Networks (WMNs) provide significant benefits over existing wirelessmulti-hop networking paradigms, offering a suitable solution for a wide range of application scenarios, spanning frompublic safety communications to community-based networks and metro scale municipal networks. Routing design is crucial to guarantee robust communication through the mesh backbone. Traditional unicast routing has shown to be ineffective when dealing with highly variable wireless channels. One of the most critical aspects is the wire- less diversity, intended as the reception of a packet at multiple forwarders, causing collisions and interference due to the broadcast nature of the wireless medium. A set of innovative routing approaches has recently been proposed as a valuable alternative to classical routing, thanks to their ability to deal with the wireless diversity as an opportunity rather than a shortcoming. The primary goal of this thesis is to deeply investigate wireless diversity-based routing inWMNs, proposing novel solutions able to significantly improve WMN performance. We extensively describe the main features of this strategy and provide a classification of themost representative solutions in literature, discussing their most relevant characteristics, advantages and disadvantages. Then, we focus on one of the most promising categories: Opportunistic Routing (OR). It exploits the multiple packet recipients offered by the wireless transmission to incrementally build a path, selecting the best next hop only after packet reception. Then, we propose a novel opportunistic routing algorithm, able to select at each hop the forwarders that maximize the throughput gain. In contrast to the common opportunistic approach, the proposed algorithm avoids any form of a priori constraint on route selection, fully leveraging all the transmission opportunities encountered during path construction. To improve its efficiency in multi-flow environments, we extend its routing strategy with an opportunistic packet scheduling algorithm and a prioritized channel access scheme, so as to facilitate the transmission of the packets that are traversing the paths providing higher performance gains. To ensure high performance in all the typical WMN application scenarios, we need to consider that in these environments channel quality may significantly vary in time and space, requiring a high degree of flexibility in the path construction process. Most of the existing solutions performlocal decisions (i.e. hop-by-hop) based on end-to-end principles. In contrast, we propose a novel routing algorithm that combines end-to-end with localized data, so as to adapt routing decisions to channel conditions at the time of packet transmission. This ensures higher reliability even in the most challenging application scenarios. The key factors determining opportunistic improvements are not clear yet, making hard to identify the conditions under which this paradigm outperforms classical unicast routing. Hence, we propose a novel routing architecture that relies on a configurable machine learning-based agent to properly select, at each node, the most suitable routing algorithm within a set of available solutions, according to network conditions and traffic characteristics. This solution represents a further step towards the definition of a wireless diversity-based routing paradigm able to ensure high performance in all WMN application scenarios.

Item Type: IMT PhD Thesis
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
PhD Course: Computer Science and Engineering
Identification Number: 10.6092/imtlucca/e-theses/30
Date Deposited: 10 Jul 2012 13:53
URI: http://e-theses.imtlucca.it/id/eprint/30

Actions (login required, only for staff repository)

View Item View Item