Subventions et des contributions :

Subvention ou bourse octroyée s'appliquant à plus d'un exercice financier. (2017-2018 à 2022-2023)

Today, web and cloud service providers offer a Faustian bargain to their customers: in exchange for service, you must give us all of your data. The advantages of cloud services over “old school” client-side software are many, including: i.) universal accessibility, ii.) share-ability, iii.) high durability of data, iv.) transparent code updates, and v.) often seemingly free service for all but premium functionality.

However, these benefits come at a large expense: loss of control of the user's own data. First and foremost, the user gives up the privacy of his data. Service providers grant themselves rights to monetize user data in their service agreements that are extremely opaque to the user. Second, the user has no means to audit the access permissions to, or retention of, the data as they are implemented by server side software. E.g., while a user may remove a photo from her home page, that same photo may be accessible via a direct URLreference long after the supposed deletion. Even more troubling, a user may simply lose access to his data due to a dispute or a change in business posture. Short of that, when a user is allowed to extract his data, he is left with an unintelligible mass without the software to navigate it. Finally, since a user often employs many cloud applications, his data is scattered across the Internet and the mechanisms for retrieving related items are few.

The research proposed here is focused on the following question. Is it possible to build applications and services that retain the advantages of the web service delivery model but improve the control users have over their own data?

We believe that the answer is yes. In my lab we are working to understand the functionality that would allow service providers to give consumers a broader range of choices in how their data is handled. In our research to date we have developed mechanisms for a web browser extension to monitor, interpose on, and provide new APIs for client-side web applications that allow for rich applications while still providing users privacy and control of their data. We propose to continue this line of research in several directions.

1.) We propose to extend our web interposition techniques to build a platform for personal web analytics. The platform will perform two essential roles. First, it will record a user’s web sessions to a secure log. Second, it will allow analysis applications access to the log while sandboxing the applications to prevent privacy leaks.

2.) We will also examine how to extend these platforms to mobile software. We will investigate whether the ARM TrustZone mechanism is sufficient to overcome the insecurities introduced by the Android OS.

3.) We will leverage the scale of existing Web services to build important security infrastructure for supporting a more diverse ecosystem of private applications such as a robust user­-to­-user public key infrastructure and an anonymity communication service as an alternative to TOR.