Tuesday, July 9, 2013 — 11:00 AM EDT

Candidate

Xiaohui Liang

Title

Security and Privacy Preservation in Mobile Social Networks

Supervisors

Shen, Sherman X. and Lin, Xiaodong (Adjunct)

Abstract

Social networking extending the social circle of people has already become an important integral part of our daily lives. As reported by ComScore, social networking sites such as Facebook and Twitter have reached 82 percent of the world's online population, representing 1.2 billion users around the world. In the meantime, fueled by the dramatic advancements of smartphones and the ubiquitous connections of Bluetooth/WiFi/3G/LTE networks, social networking further becomes available for mobile users and keeps them posted on the up-to-date worldwide news and messages from their friends and families anytime anywhere. The convergence of social networking, advanced smartphones, and stable network infrastructures brings us a pervasive and omnipotent communication platform, named mobile social network (MSN), helping us stay connected better than ever. In the MSN, multiple communication techniques help users to launch a variety of applications in multiple communication domains including single-user domain, two-user domain, user-chain domain, and user-star domain. Within different communication domains, promising mobile applications are fostered. For example, nearby friend search application can be launched in the two-user or user-chain domains to help a user find other physically-close peers who have similar interests and preferences; local service providers disseminate advertising information to nearby users in the user-star domain; and health monitoring enables users to check the physiological signals in the single-user domain.

Despite the tremendous benefits brought by the MSN, it still faces many technique challenges among of which security and privacy protections are the most important ones as smartphones are vulnerable to security attacks, users easily neglect their privacy preservation, and mutual trust relationships are difficult to be established in the MSN. In this thesis, we explore the unique characteristics and study typical research issues of the MSN. We conduct our research with a focus on security and privacy preservation while considering human factors. Specifically, we consider the profile matching application in the two-user domain, the cooperative data forwarding in the user-chain domain, the trustworthy service evaluation application in the user-star domain, and the healthcare monitoring application in the single-user domain. The main contributions are, i) considering the human comparison behavior and privacy requirements, we first propose a novel family of comparison-based privacy-preserving profile matching (PPM) protocols.

The proposed protocols enable two users to obtain comparison results of attribute values in their profiles, while the attribute values are not disclosed. Taking user anonymity requirement as an evaluation metric, we analyze the anonymity protection of the proposed protocols. From the analysis, we found that the more comparison results are disclosed, the less anonymity protection is achieved by the protocol. Further, we explore the pseudonym strategy and an anonymity enhancing technique where users could be self-aware of the anonymity risk level and take appropriate actions when needed; ii) considering the inherent MSN nature --- opportunistic networking, we propose a cooperative privacy-preserving data forwarding (PDF) protocol to help users forward data to other users. We indicate that privacy and effective data forwarding are two conflicting goals: the cooperative data forwarding could be severely interrupted or even disabled when the privacy preservation of users is applied, because without sharing personal information users become unrecognizable to each other and the social interactions are no longer traceable. We explore the morality model of users from classic social theory, and use game-theoretic approach to obtain the optimal data forwarding strategy.

Through simulation results, we show that the proposed cooperative data strategy can achieve both the privacy preservation and the forwarding efficiency; iii) to establish the trust relationship in a distributed MSN is a challenging task. We propose a trustworthy service evaluation (TSE) system, to help users exchange their service reviews toward local vendors. However, vendors and users could be the potential attackers aiming to disrupt the TSE system. We then consider the review attacks, i.e., vendors reject and modify the authentic reviews of users, and the Sybil attacks, i.e., users abuse their pseudonyms to generate fake reviews. To prevent these attacks, we explore the token technique, the aggregate signature, and the secret sharing techniques. Simulation results show the effectiveness of the TSE system can be guaranteed; iv) to improve the efficiency and reliability of communications in the single-user domain, we propose a prediction-based secure and reliable routing framework (PSR). It can be integrated with any specific routing protocol to improve the latter's reliability and prevent data injection attacks during data communication. We show that the regularity of body gesture can be learned and applied by body sensors such that the route with the highest predicted link quality can always be chose for data forwarding. The security analysis and simulation results show that the PSR significantly increases routing efficiency and reliability with or without the data injection attacks.

Location 
EIT building
Room 4152

,

S M T W T F S
26
27
28
29
30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1
2
3
4
5
6
  1. 2020 (130)
    1. September (1)
    2. August (1)
    3. July (14)
    4. June (17)
    5. May (29)
    6. April (27)
    7. March (13)
    8. February (20)
    9. January (16)
  2. 2019 (282)
    1. December (16)
    2. November (32)
    3. October (19)
    4. September (26)
    5. August (26)
    6. July (40)
    7. June (24)
    8. May (23)
    9. April (35)
    10. March (25)
    11. February (9)
    12. January (10)
  3. 2018 (150)
  4. 2017 (212)
  5. 2016 (242)
  6. 2015 (242)
  7. 2014 (268)
  8. 2013 (192)
  9. 2012 (31)