This research aims at exploring novel approaches that can enable information and communications technologies to play significant roles in aiding the development agenda in challenged rural and poor environments, focusing on ad-hoc and sensor networks. This is addressed through two stages: first the needs and particularities of a typical, rural farming community in Sub-Saharan Africa are identified and analyzed through work carried out within the Engineering and Physical Sciences project VESEL (village e-science for life). Based on this higher level treatment the thesis employs an integrated approach to Information and Communication Technologies for Development (lCTD) and proposes an informed top-bottom generic technical framework for utilizing WSNs as a development enabling tool in challenged communications contexts. Second the thesis extracts specific low level technical challenges and focuses on how joint distributed storage management, rate-adaptive data generation and topology control can enable ad-hoc and sensor j networks operate efficiently and reliably in extreme conditions where network disconnections and interruptions are the norm. This results in developing new distributed algorithms for managing the network's storage and data generation in a way that adaptively responds to the network state and application requirements. These show considerable improvement over conventional 'local' storage approaches while keeping their advantage under different deployment conditions. The thesis also studies the spatio-temporal network devolution and connectivity- related effects of radio-wave shadowing and fading and quantifies their impacts on the performance of distributed storage and data management. By utilizing network optimization tools the thesis further shows how the problem of distributed storage in challenged sensor networks can be treated as a distributed storage allocation problem where effective strategies can be established to jointly optimize storage selection, storage routing and storage energy consumption. The design of optimum storage topologies is also possible through this approach. This part results in designing practical decentralized heuristics that achieve close-to-optimum dynamic allocation of distributed storage and hence enable storage-aided efficient data dissemination in challenged sensor networks lacking a continuous data upload capability.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:555864 |
| Date | January 2011 |
| Creators | Kabashi, Amar Hussein |
| Publisher | University of Leeds |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0112 seconds