Privacy Aware Architecture for Smart City Ecosystems

  • Ali S. Abuzinjal

Student thesis: Master's Thesis

Abstract

In future smart cities, the exchange of data will play a vital role to create cognitive cities that need to react to critical fluctuations in a timely manner. According to recent study, 75% of available data today are personal data generated by a myriad of devices and systems. By aggregating and analyzing the generated data, governments can create a knowledge base to enhance dwellers life. While the just cause is justifiable, the entailed risk of sharing the data beyond their operational boundaries imposes a privacy risk to data subjects. Historically, privacy practitioners have focused their effort to solve this issue based on their temporal assumptions. With the magnitude of data, driven by the digital transformation, new approaches should be investigated to manage privacy risks arising from the demand of connected and open data strategies. The existing privacy solutions are mainly focused on single release, without considering the release environment of the data, which fall short when adversaries manage to link existing datasets by exploiting different data attributes and hence disclose the data subject identity. In our research, we present a novel privacy-aware framework to control data releases and annotate the data by a semantic metadata vocabulary relaying on W3C semantic web standards, to orchestrate the dissemination of data and mitigate the risk of data linkage attacks on data subjects. The framework implements a data-dependent approach and applies the necessary anonymization techniques to preserve subjects identities and ensure the compliance of the data releases, in the case of sequential queries, to the policies imposed by the data owners.
Date of AwardMay 2020
Original languageAmerican English
SupervisorRabeb Mizouni (Supervisor)

Keywords

  • Smart Cities
  • Digital Transformation
  • Privacy Risks
  • Privacy-Aware Framework
  • Data-dependent Approach
  • Semantic Web
  • Linked Open Data (LOD)
  • QIDs.

Cite this

'