Social-Aware Data Dissemination in Opportunistic Mobile Networks

Open Access
- Author:
- Gao, Wei
- Graduate Program:
- Computer Science and Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- May 17, 2012
- Committee Members:
- Guohong Cao, Dissertation Advisor/Co-Advisor
C Lee Giles, Committee Member
George Kesidis, Committee Member
Xiaolong Zhang, Committee Member - Keywords:
- opportunistic mobile networks
data dissemination
social networks
mobile computing
user interest
centrality
community - Abstract:
- Recent technical advance of personal hand-held mobile computing devices necessities mobile data accessibility, which realizes "anytime, anywhere" computing without being constrained by the wired network links and fixed network locations. Such data accessibility could be realized by exploring unused communication opportunities between the short-range radios, such as Bluetooth or WiFi, of mobile devices, when they opportunistically move into the communication range of each other. These devices, therefore, form Opportunistic Mobile Networks. The specific goal of this dissertation is to develop comprehensive solutions for providing mobile data accessibility in opportunistic mobile networks, when the cellular 3G or WiFi network infrastructure is unavailable. In particular, this dissertation explores the social behavior patterns of mobile users to address the challenges in network uncertainty, low quality, and subjectivity, and ensures that data can be efficiently disseminated to users being interested in the data by selecting appropriate nodes as relays to carry and forward data. An overall research strategy is to explore the correspondence of sociological concepts, including centrality, community, and homophily, in opportunistic communication among mobile users. Analytical measures of sociological concepts are devised to extract users' implicit social links from observations of their social behavior patterns. These measures are then exploited for efficient network decisions on data dissemination. This dissertation synthesizes three aspects of users' social behavior patterns for prompt and efficient data dissemination. First, we investigate users' social contact patterns, from which we propose centrality and community measures for facilitating opportunistic communication. Second, we take the satisfaction of socially correlated user interest into consideration, and incorporate users' social interests into measuring user centrality. Third, we study the impact of homophily on data dissemination performance, and provide theoretical analysis on the optimal relay selection policies, from the correlation between friendship and interest similarity of mobile users.