Towards context aware opportunistic forwarding in social pervasive systems

Abstract

Recent advances in mobile device sensor technology, coupled with a wealth of structured and accessible data from social networks, have together formed a data-rich ecosystem. Such an ecosystem is very wealthy in a bi-directional context that can flow between the mobile and social worlds in order to promote the creation of an elitist breed of pervasive services and applications. We label the breed resulting from the merger as Social Pervasive Systems (SPS). We review literature of the domains of social networks and mobile pervasive systems to study prior research attempts to merge both domains as detailed in Chapter 2. We begin by presenting our observations in a timeline that illustrates the progress of the merger attempts. From this study, we are able to identify a collection of services and application families that can rise as a byproduct of the merger. We also identify a set of challenges that deter the formation of systems of this kind and propose solutions for them. Although the internet access is pervasive and ubiquitous in the developed countries, it is scarce in the developing and the undeveloped economies. With the current setup in the developing countries where users own smart devices and demand access to the internet, but suffer from the poor network infrastructure, there rises the need for alternative network connectivity such as delay tolerant networks (DTNs) and opportunistic networks. Alternative technologies have been used to compensate for the scarceness of the network infrastructure and the network disconnection. In this research, we focus on a subset of the SPS applications; namely, the social-based opportunistic forwarding algorithms that are highly recommended in the domain of areas with challenged network infrastructure coinciding with pervasive mobile usage and high demand for internet access and connectivity. We focus on the challenges facing such algorithms and the drawbacks in performance as relates to efficiency, effectiveness, power awareness, and utilization fairness. From there, we propose and experiment with solutions to improve the performance of opportunistic forwarding algorithms that are much needed in environments which lack network infrastructure or those that are vulnerable to frequent disruptions. These solutions employ bi-directional context from the mobile and social worlds pertaining to user mobility, social interest, power awareness, and contact durations. Four major contributions are proposed in this research. The first and second contributions demonstrate an improvement over existing popular opportunistic forwarding algorithms, such as the People Rank algorithm, the Socialcast algorithm, and the Sensor Context-Aware Routing protocol (SCAR) by integrating interest awareness and power awareness into these algorithms. We propose the PI-SOFA framework as a framework for integrating interest and power awareness into social-aware opportunistic forwarding algorithms as detailed in Chapter 3; PI-SOFA integration implemented versions are described in detail in Chapter 5. We question the accuracy of Space syntax metrics in defining the attraction points in a given urban area and argue that this negatively affects the performance and the accuracy of forwarding decisions. This is the third proposed contribution which is presented in Section 3.2 and its proposed implemented versions are described in detail in Chapter 6. The fourth proposed contribution is proposing dynamic adaptive ranking that dynamically changes the weight of the factors controlling the node's rank based on the current context. Details of the dynamic adaptive ranking are illustrated in Section 3.3, and its implemented versions are described in Chapter 7. All our contributions are empirically evaluated via our proposed simulator SAROS, our fifth contribution, which is presented in detail in Chapter 4. Throughout our research, the simulations conducted with SAROS utilize imported datasets that include both realistic and synthesized mobility traces, social profiles, social relationships, power consumption models, as well as data that are generated by the simulator itself. Detailed description of the used or generated datasets is presented in Section 4.2. The evaluation metrics that are used in the conducted experiments, along with the utilized scientific methodology are also provided. Finally, statistical analysis is conducted to produce the recommended regression model of the six main performance metrics of the dynamic adaptive ranking approach which is detailed in Section 7.7.

Department

Computer Science & Engineering Department

Graduation Date

2-1-2017

Submission Date

December 2016

First Advisor

Aly, Sherif

Second Advisor

Harras, Khaled

Committee Member 1

Shalan, Mohamed

Committee Member 2

Kassas, Sherif

Extent

293 p.

Document Type

Doctoral Dissertation

Rights

The author retains all rights with regard to copyright. The author certifies that written permission from the owner(s) of third-party copyrighted matter included in the thesis, dissertation, paper, or record of study has been obtained. The author further certifies that IRB approval has been obtained for this thesis, or that IRB approval is not necessary for this thesis. Insofar as this thesis, dissertation, paper, or record of study is an educational record as defined in the Family Educational Rights and Privacy Act (FERPA) (20 USC 1232g), the author has granted consent to disclosure of it to anyone who requests a copy.

Institutional Review Board (IRB) Approval

Approval has been obtained for this item

Comments

I have to acknowledge Mr. Youssef Jameel to granting me the PhD fellowship at AUC for 3 and a half Year. May God rewards him for all his good deeds. Also, I have full due gratitude to my dear parents for supporting me and encouraging to go on in my studies since the first days of my life. Finally, I am grateful for my advisors; Dr. Sherif Aly and Dr. Khaled Harras, for their continuous support and for the knowledge they conveyed to me along 6 years of research.

This document is currently not available here.

Share

COinS