Inferring social relationships from technology-level device connections

Abstract

Technology is present in every area of our lives and, for many, life without it has become unthinkable. As a consequence of this dependence and the extent to which technology devices (computers, tablets and smartphones) are being used for work and social activities, a clear coupling between devices and their owners can now be observed. By coupling, we specifically refer to the fact that information present on a person's device, be it user-generated or created by the native OS, can produce great insight into their life. In this paper, we look to exploit this coupling to investigate whether connections between technology devices recorded in system log-files, can be used to make inferences about the social relationships between device owners. A key motivation here is to better understand and elucidate the privacy risks associated with the digital footprints that we as humans (often inadvertently) create. Our work draws upon Social Network Analysis and basic Computer Forensics to develop and achieve the inference goals. From our preliminary experimentation, we demonstrate that human social relationships can indeed be inferred even within our limited initial scope. To further investigate the level of privacy exposure from technology-level links, we outline a more comprehensive plan of experimentation that will be conducted in future work.

Exploring the Risks to Identity Security and Privacy in Cyberspace

“…An analysis of the scanned log data revealed several facts worthy of the attention of all who use the aforementioned devices. We present only a few of our findings here [7].…”

“…The nodes are colorcoded with scanned systems labeled according to their computer type and shown in red and yellow and discovered systems presented in other colors. Graph edges represent connections between the devices (for example, a USB plugged into a system, or a system connected to a Wi-Fi access point) [7]. of them are currently available.…”

Identifying Key-Players in Online Activist Groups on the Facebook Social Network

“…This may allow us to identify hidden friendships, where users try to cover their friendship by not creating a direct link between them. This type of analysis is valuable to the field of cybercrime investigation, particular in combination with work such as [38], [39], where we do not have enough information about the group hierarchy, however, may have access to their communication.…”

Behind the scenes: a cross-country study into third-party website referencing and the online advertising ecosystem

“…The first level sought to apply social network analysis (SNA) techniques to build a visual network that would better allow the identification of relationships within the site-domain listings. SNA can be described as a set of methods that allow the links between elements (e.g., people, devices, or things) to be studied [30]. We were keen on applying this technique to identify the third-party domains that were the most common across sites generally, or specifically when considering the site's locale.…”