Skip to Main content Skip to Navigation

Unveiling and Controlling Online Tracking

Jagdish Prasad Achara 1, 2 
2 PRIVATICS - Privacy Models, Architectures and Tools for the Information Society
Inria Grenoble - Rhône-Alpes, CITI - CITI Centre of Innovation in Telecommunications and Integration of services, Inria Lyon
Abstract : It is no surprise, given smartphones convenience and utility, to see their wide adoption worldwide. Smartphones are naturally gathering a lot of personal information as the user communicates, browses the web and runs various Apps. They are equipped with GPS, NFC and digital camera facilities and therefore smartphones generate new personal information as they are used. Since they are almost always connected to the Internet, and are barely turned off, they can potentially reveal a lot of information about the activities of their owners. The close arrival of smart-­‐watches and smart-­‐glasses will just increase the amount of personal information available and the privacy leakage risks. This subject is closely related to the Mobilitics project that is currently conducted by Inria/Privatics and CNIL, the French data protection authority [1][2][3]. Therefore, the candidate will benefit from the investigations that are on progress in this context, in order to understand the situation and the trends. The candidate will also benefit from all the logging and analysis tools we developed for the iOS and Android Mobile OSes, as well as the experienced gained on the subject. Another question is the arrival of HTML5 based Mobile OSes, like Firefox OS: it clearly opens new directions as it "uses completely open standards and there’s no proprietary software or technology involved" (Andreas Gal, Mozilla). But what are the implications from a Mobile OS privacy point of view? That's an important topic to analyze. Beyond understanding the situation, the candidate will also explore several directions in order to improve the privacy control of mobile devices. First of all, a privacy-­‐by-­‐design approach, when feasible, is an excellent way to tackle the problem. For instance the current trend is to rely more and more on cloud-­‐based services, either directly (e.g., via Dropbox, Instagram, Social Networks, or similar services), or indirectly (e.g., when a backup of the contact, calendar, accounts databases is needed). But pushing data on cloud-­‐based systems, somewhere on the Internet, is in total contradiction with our privacy considerations. Therefore, an idea is to analyze and experiment with personal cloud services (e.g., ownCLoud, diaspora) that are fully managed by the user. Here the goal is to understand the possibilities, the opportunities, and the usability of such systems, either as a replacement or in association with commercial cloud services. Another direction is to carry out behavioral analyses. Indeed, in order to precisely control the privacy aspects, at one extreme, the user may have to deeply interact with the device (e.g., through pop-ups each time a potential privacy leak is identified), which negatively impacts the usability of the device. At the other extreme, the privacy control may be oversimplified, in the hope not to interfere too much with the user, as is the case with the Android static authorizations or the one-­‐time pop-­‐ups of iOS6. This is not appropriate either, since using private information once is not comparable to using it every minute. A better approach could be to perform, with the help of a machine learning system for instance, a dynamic analysis of the Mobile OS or App behavior from a privacy perspective and to interfere with the user only when it is deemed appropriate. This could enable a good tradeoff between privacy control and usability, with user actions only when meaningful. How far such a behavioral analysis can go and what are the limitations of the approach (e.g., either from a CPU/battery drain perspective, or in front of programming tricks to escape the analysis) are open questions. Tainting techniques applied to Mobile OSes (e.g., Taint-­Droid) can be used as a basic bloc to build a behavioral analysis tool, but they have limited accuracy are unable to analyze native code and have poor performances.
Document type :
Complete list of metadata

Cited literature [152 references]  Display  Hide  Download
Contributor : ABES STAR :  Contact
Submitted on : Wednesday, January 10, 2018 - 5:00:11 PM
Last modification on : Thursday, August 4, 2022 - 5:18:38 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01386405, version 2


Jagdish Prasad Achara. Unveiling and Controlling Online Tracking. Mobile Computing. Université Grenoble Alpes, 2016. English. ⟨NNT : 2016GREAM069⟩. ⟨tel-01386405v2⟩



Record views


Files downloads