UBC Theses and Dissertations
StrongHold : a secure data platform for smart homes Clapauch, Jaques
Security is an important concern in day to day life, far augmented when related to sensitive data, such as private information. As such, it is paramount to protect environments and devices wherein the greater amount of information is private, such as in a home, and its encompassing devices; especially so, as the level of privacy desired in a home is much greater than other private mediums. The issue is further exacerbated in a smart home scenario, where applications are ubiquitous and are constantly communicating with the outside world: this impending technological innovation results in previously private data suddenly becoming accessible through non-physical means, allowing potential breaches in privacy. The accessibility to previously unreachable means brings forth new threats that must be tackled to ensure proper confidentiality is kept. Furthermore, proper accessibility of the physical devices must be engaged, due to their newfound non-physically accessibly nature. We survey the security threats existent in smart home environments, along with possible solutions to mitigate the threats. We use methods ranging from human computer interaction, storage optimization, and behavioral learning to better ensure proper functionality. We integrate the solutions in a cohesive form to be applied to any smart home environment that wishes to best keep high confidentiality, availability and integrity. We test the system to ensure the secure data messaging platform provides sufficient throughput for high definition media storage in real time, as potentially necessary in a smart home We supplement our study by creating a system capable of functioning on top of the smart server, which unobtrusively automates the daily task of recipe selection, via meal advocation through past meal selection; we do so unobtrusively in an attempt to prevent deviation from conventional home environments. We present a method which possesses higher granularity than the present meal recommendation technologies, and yet, requires less interaction than the web equivalents. Finally, we interface the algorithm to a real smart home server, evaluate the application on real users, and assess their reaction to the technology.
Item Citations and Data
Attribution 2.5 Canada