Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.680700
Title: Mobile social networking aided content dissemination in heterogeneous networks
Author: Hu, Jie
ISNI:       0000 0004 5916 7251
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 2015
Availability of Full Text:
Access through EThOS:
Full text unavailable from EThOS. Thesis embargoed until 22 Oct 2018
Access through Institution:
Abstract:
Thanks to the rapid development of the wireless Internet, numerous mobile applications provide platforms for Mobile Users (MUs) to share any Information of Common Interest (IoCI) with their friends. For example, mobile applications of Facebook and Twitter enableMUs to share information via posts and status updates. Similarly, the mobile application of Wase (http://www.waze.com/) enables the drivers to share the real-time traffic information collected by themselves. However, at the time of writing, the dissemination process of the IoCI is predominantly supported by the Centralised Infrastructure (CI) based communication. In order to receive the IoCI, the mobile devices of MUs have to be connected either to a Base Station (BS) or to a Wi-Fi hotspot. However, the CI-based information dissemination faces the following three limitations: i) intermittent connectivity in rural areas; ii) overloaded CI-based network; iii) inefficient data service in densely populated areas. Due to the rapid development of the powerful mobile computing technique, mobile devices are typically equipped with large storage capacity, say dozens or possibly hundreds of Giga Bytes. Furthermore, they support multiple communication standards, such as Infrared, Bluetooth and Wi-Fi modules, in support of direct peer-to-peer communications. As a result, this treatise contributed towards mitigating the above-mentioned design problems in the conventional CI-based information dissemination by seeking assistance both from the opportunistic contacts and opportunistic multicast amongst MUs, who share common interest in the same information. This results in an integrated cellular and opportunistic network. Since mobile devices are carried by MUs, exploring the social behaviours exhibited by individuals and the social relationships amongst MUs may assist us in enhancing the communication experience. We firstly study an integrated cellular and large-scale opportunistic network, where theMUs are sparsely distributed within a large area. Due to the large size of the area and the sparse distribution of the MUs, the connectivity between a BS and a MU as well as that between a pair of MUs exhibit an intermittent nature. As a result, the information delivery has to be realised by the opportunistic contact between a transmitter and receiver pair, when the receiver enter the range of the transmitter. However, successful information delivery requires that the duration of this opportunistic contact has to be longer than the downloading period of the information. This integrated network is relied upon for disseminating delay-tolerant IoCI amongst the MUs. We model the information dissemination in this integrated network by a Continuous-Time-Pure-Birth-Markov-Chain (CT-PBMC) and further derive its relevant delay metrics and information delivery ratio. With the assistance of large-scale opportunistic networks, the information delivery ratio before the IoCI expires may be doubled, when compared to the conventional CI-based information dissemination. Furthermore, upon modelling the contact history of the MUs as a social network, social centrality based schemes are proposed for the sake of off-loading tele-traffic from the potentially congested CI to the large-scale opportunistic network. As demonstrated by our simulation results, in the scenario considered, as many as 58% MUs can be served by the large-scale opportunistic network before the oCI expires. In the above-mentioned large-scale networks, the MUs tend to be dispersed. By contrast, in the densely populated scenario, where numerous MUs can be found within a small area, classic BS-aided multicast is often invoked as a traditional measure of disseminating the IoCI by relying on the broadcast nature of the wireless channels. However, BS-aided multicasting becomes inefficient, when the number of MUs is high. If we efficiently exploit the redundant copies of the IoCI held by the already served MUs and activate these IoCI holders as potential relays for the next stage of cooperative multicast, the resultant diversity gain beneficially accelerates the information dissemination process. This approach may be regarded as opportunistic cooperative multicast, since no deterministic relay selection scheme is required. Its promising advantages experienced during disseminating the delay-sensitive IoCI across the densely populated area considered motivate us to study an integrated cellular and small-scale opportunistic network. By jointly considering both the effects of the channel model in the physical layer, as well as the resource scheduling in the Medium Access-Control (MAC) layer and the information dissemination protocol in the network layer, we model the information dissemination process by a Discrete-Time-Pure-Birth-Markov-Chain (DT PBMC). Apart from the above-mentioned factors related to wireless transmission, we also consider various MUs’ social characteristics, such as their altruistic behaviours and their geographic social relationships. Relying upon the above-mentioned DT-PBMC, we are capable of studying the delay versus energy dissipation trade-off during the information dissemination process. As demonstrated by our simulation results, the integrated cellular and small-scale opportunistic network considered is capable of substantially reducing the total information dissemination delay and the total energy dissipation of the classic BS-aided multicast. However, these benefits are achieved at the cost of the additional energy dissipated by the individual MUs. In order to further reduce both the information dissemination delay and the energy dissipation, Social Network Analysis (SNA) tools are relied upon for proposing a range of efficient resource scheduling approaches in the MAC layer. As demonstrated by our simulation results, the so-called shortest-shortest-distance scheduling regime outperforms its counterparts in terms of both its delay and energy metrics. Since the distance-related path-loss predetermines the successful information delivery in the scenario of the integrated cellular and small-scale opportunistic network, it is crucial for us to study the statistical properties of the random distance between a transmitter and receiver pair. As a result, we derive the closedform distributions of both the random distance between a pair of MUs and that between a BS and MU pair for different scenarios. Apart from assisting us in analysing the information dissemination process, these results may be further relied upon for evaluating the path loss, the throughput, the spectral efficiency and the outage probability in mobile networks. Although mobile communication techniques evolved from the well known analog ‘1G’ mobile networks to the emerging heterogeneous ‘5G’ mobile networks, the operational systems still rely on the CI-dominated ‘Comm 1.0’ era. Explicitly, direct interaction amongst the MUs without any aid of the CI is rare. Since powerfulmobile devices and pervasive social networking services are popular amongst theMUs, more direct interaction amongst theMUs would be advocated for the sake of achieving a more reliable, more prompt, and more energy-efficient communication experience. This treatise may contribute in a modest way towards the new ‘Comm 2.0’ era and may inspire further efforts from both the industrial and academic communities so as to embrace both the opportunities and the challenges of this new era from both the technical and economic perspectives.
Supervisor: Hanzo, Lajos Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.680700  DOI: Not available
Keywords: QA75 Electronic computers. Computer science
Share: