INTACT: Storing unbounded data streams on mobile devices to unlock user privacy at the edge
Keywords: mobile, storage, location, privacy
TL;DR: INTACT is a middleware solution to efficiently store time series on memory-constrained devices. Thanks to INTACT, clustering algorithms used by points-of-interest attacks can run at the edge, hence mitigating privacy disclosure threats.
Abstract: Mobile devices are producing larger and larger data streams, such as location streams, which are consumed by machine learning pipelines to deliver location-based services to end users. Such data streams are generally uploaded and centralized to be processed by third parties, potentially exposing sensitive personal information. In this context, existing protection mechanisms, such as Location Privacy Protection Mechanisms (LPPMs), have been investigated. Alas, none of them have effectively been implemented, nor deployed in mobile devices to enforce user privacy at the edge of a network. We believe that the effective deployment of LPPMs on mobile devices faces a major challenge: the storage of unbounded data streams.
This article introduces INTACT, a cross-platform framework that leverages a piece-wise linear approximation technique, dubbed FLI, to increase the storage capacity of mobile devices. Then, we combine this storage capability with Divide & Stay, a new privacy preservation technique to execute Points of Interest (POIs) inference. By enabling in situ POIs inference, the sensitivity of location streams can be assessed to better enforce user privacy. Finally, we deploy all INTACT components on Android and iOS to demonstrate that a real deployment of LPPMs on mobile phones is now possible.
Area: Wireless Embedded Systems
Type: Solution
Revision: No
Submission Number: 5
Loading