Li, FengPeng, WeiZou, XukaiLi, Ninghui2016-01-072016-01-072015https://hdl.handle.net/1805/7995http://dx.doi.org/10.7912/C2/2311Indiana University-Purdue University Indianapolis (IUPUI)A defining characteristic of smartphones is the availability of short-range radio transceivers (the proximate channel) such as Bluetooth, NFC, and Wi-Fi Direct, in addition to traditional long-range cellular telecommunication technologies (the cellular channel). Coupled with smartphones’ portability and their human users’ mobility, the proximate channel provides opportunistic proximate links as a supplement/alternative to the cellular channel’s persistent infrastructural links for data communication. Opportunistic proximate links have a diverse set of applications, with each application scenario bringing a unique set of often conflicting objectives to balance. This dissertation presents a study on several problems regarding the application of opportunistic proximate links in smartphone networks. The first part of this dissertation, which includes Chapter 2, 3, and 4, focuses on the cost-effective distribution of content using opportunistic proximate links, and examines several applications: 1. Chapter 2 is on the use of opportunistic proximate links in selecting a representative subset from a set of smartphones for prioritized defense deployment in a Bring-Your-Own-Device (BYOD) enterprise network environment. 2. Chapter 3 is on the use of opportunistic proximate links for offloading bounded-delay-tolerant topical content from cellular persistent infrastructural links. 3. Chapter 4 is on the use of opportunistic proximate links in a generalized scenario of content distribution in a smartphone network that is heterogeneous in the availability of cellular persistent infrastructural links. The second part of this dissertation, which includes Chapter 5 and 6, considers the opposite problem of preventing the distribution of unwanted content (mobile malware) over opportunistic proximate links and the supplementary problem of detecting mobile malware. Chapter 5 considers a probabilistic behavioral malware detection framework for delay-tolerant smartphone networks that are connected by opportunistic proximate links. Solutions to several challenging problems that are unique to decentralized and opportunistic nature of such networks, including “balance between insufficient evidence and evidence collection risk,” “liars,” and “defectors” are proposed and evaluated. Based on the widely used Android mobile computing platform, Chapter 6 presents the design, implementation, and evaluation of a novel declarative approach to static binary analysis of Android apps, which underlies the problem of detecting malware on the Android platform. Real Android malware samples are analyzed, and techniques to robustly handle them are proposed and evaluated.en-USAttribution-ShareAlike 3.0 United StatesThesis