Context-Centric Security

TitleContext-Centric Security
Publication TypeConference Paper
Year of Publication2012
AuthorsTiwari, M., Mohan P., Osheroff A., Alkaff H., Shi E., Love E., Song D., & Asanović K.
Other Numbers3456

Users today are unable to use the rich collection ofthird-party untrusted applications without risking significant privacyleaks. In this paper, we argue that current and proposed applicationsand data-centric security policies do not map well to users’ expectationsof privacy. In the eyes of a user, applications and peripheraldevices exist merely to provide functionality and should haveno place in controlling privacy. Moreover, most users cannot handleintricate security policies dealing with system concepts such aslabeling of data, application permissions and virtual machines. Notonly are current policies impenetrable to most users, they also leadto security problems such as privilege-escalation attacks and implicitinformation leaks.

Our key insight is that users naturally associate data with realworldevents, and want to control access at the level of human contacts.We introduce Bubbles, a context-centric security system thatexplicitly captures user’s privacy desires by allowing human contactlists to control access to data clustered by real-world events. Bubblesinfers information-flow rules from these simple context-centricaccess-control rules to enable secure use of untrusted applicationson users’ data.

We also introduce a new programming model for untrusted applicationsthat allows them to be functional while still upholding theusers’ privacy policies. We evaluate the model’s usability by portingan existing medical application and writing a calendar app fromscratch. Finally, we show the design of our system prototype runningon Android that uses bubbles to automatically infer all dangerouspermissions without any user intervention. Bubbles preventsAndroid-style permission escalation attacks without requiring usersto specify complex information flow rules.


The authors would like to thank the anonymous reviewersfor providing useful comments on this paper. This materialis based upon work supported by the AFOSR under MURIaward FA9550-09-1-0539, by the National Science Foundationunder Grant #1136996 to the Computing Research Associationfor the CIFellows Project, by NSF awards CPS-0932209 and CPS-0931843, and by Intel through ISTC forSecure Computing.

Bibliographic Notes

Proceedings of the 7th USENIX Workshop on Hot Topics in Security (HotSec'12), Bellevue, Washington

Abbreviated Authors

M. Tiwari, P. Mohan, A. Osheroff, H. Alkaff, E. Shi, E. Love, D. Song, and K. Asanovi?

ICSI Research Group


ICSI Publication Type

Article in conference proceedings