Theses and Dissertations

Author

Arun Thapa

Issuing Body

Mississippi State University

Advisor

Li, Pan

Committee Member

Topsakal, Erdem

Committee Member

Fowler, James E.

Committee Member

Du, Qian (Jenny)

Date of Degree

1-1-2014

Document Type

Dissertation - Open Access

Degree Name

Doctor of Philosophy

College

James Worth Bagley College of Engineering

Department

Department of Electrical and Computer Engineering

Abstract

The explosive growth of Online Social Networks (OSNs) over the past few years has redefined the way people interact with existing friends and especially make new friends. OSNs have also become a great new marketplace for trade among the users. However, the associated privacy risks make users vulnerable to severe privacy threats. In this dissertation, we design protocols for private distributed social proximity matching and a private distributed auction based marketplace framework for OSNs. In particular, an OSN user looks for matching profile attributes when trying to broaden his/her social circle. However, revealing private attributes is a potential privacy threat. Distributed private profile matching in OSNs mainly involves using cryptographic tools to compute profile attributes matching privately such that no participating user knows more than the common profile attributes. In this work, we define a new asymmetric distributed social proximity measure between two users in an OSN by taking into account the weighted profile attributes (communities) of the users and that of their friends’. For users with different privacy requirements, we design three private proximity matching protocols with increasing privacy levels. Our protocol with highest privacy level ensures that each user’s proximity threshold is satisfied before revealing any matching information. The use of e-commerce has exploded in the last decade along with the associated security and privacy risks. Frequent security breaches in the e-commerce service providers’ centralized servers compromise consumers’ sensitive private and financial information. Besides, a consumer’s purchase history stored in those servers can be used to reconstruct the consumer’s profile and for a variety of other privacy intrusive purposes like directed marketing. To this end, we propose a secure and private distributed auction framework called SPA, based on decentralized online social networks (DOSNs) for the first time in the literature. The participants in SPA require no trust among each other, trade anonymously, and the security and privacy of the auction is guaranteed. The efficiency, in terms of communication and computation, of proposed private auction protocol is at least an order of magnitude better than existing distributed private auction protocols and is suitable for marketplace with large number of participants.

URI

https://hdl.handle.net/11668/21046

Comments

asymmetric social proximity||private auction protocols||online social networks||private matching protocols

Share

COinS