Jefferson's Wheel

Our paper, Privacy Protection for Social Networking Platforms by Adrienne Felt and David Evans is now available [PDF]. Adrienne Felt will present the paper at the Web 2.0 Security and Privacy 2008 (in conjunction with 2008 IEEE Symposium on Security and Privacy) in Oakland, CA on May 22, 2008.

Abstract

Social networking platforms integrate third-party content into social networking sites and give third-party developers access to user data. These open interfaces enable popular site enhancements but pose serious privacy risks by exposing user data to third-party developers. We address the privacy risks associated with social networking APIs by presenting a privacy-by-proxy design for a privacy-preserving API. Our design is motivated by an analysis of the data needs and uses of Facebook applications. We studied 150 popular Facebook applications and found that nearly all applications could maintain their functionality using a limited interface that only provides access to an anonymized social graph and placeholders for user data. Since the platform host can control the third party applications’ output, privacy-by-proxy can be accomplished by using new tags and data transformations without major changes to either the platform architecture or applications.

Full paper (8 pages): [PDF]
Project Website

[Added 25 May]: Talk slides (by Adrienne Felt): [PDF]

Our paper, Hiding in Groups: On the Expressiveness of Privacy Distributions by Karsten Nohl and David Evans, is now available: PDF (15 pages). Karsten Nohl will present the paper at the 23^rd International Information Security Conference (SEC 2008, Co-located with IFIP World Computer Congress 2008) in Milan, Italy, 8-10 September 2008.

Abstract

Many applications inherently disclose information because perfect privacy protection is prohibitively expensive. RFID tags, for example, cannot be equipped with the cryptographic primitives needed to completely shield their information from unauthorized reads. All known privacy protocols that scale to the anticipated sizes of RFID systems achieve at most modest levels of protection. Previous analyses found the protocols to have weak privacy, but relied on simplifying attacker models and did not provide insights into how to improve privacy. We introduce a new general way to model privacy through probability distributions, that capture how much information is leaked by different users of a system. We use this metric to examine information leakage for an RFID tag from the a scalable privacy protocol and from a timing side channel that is observable through the tag’s random number generator. To increase the privacy of the protocol, we combine our results with a new model for rational attackers to derive the overall value of an attack. This attacker model is also based on distributions and integrates seamlessly into our framework for information leakage. Our analysis points to a new parameterization for the privacy protocol that significantly improves privacy by decreasing the expected attack value while maintaining reasonable scalability at acceptable cost.

Full paper (15 pages): [PDF]

Extended Technical Report (18 pages): [PDF]

Our paper on supporting untrusted content in aggregated web pages is now available:

• Adrienne Felt, Pieter Hooimeijer, David Evans, Westley Weimer. Talking to Strangers Without Taking Their Candy: Isolating Proxied Content. First International Workshop on Social Network Systems, Glasgow, Scotland, April 2008. (PDF, 6 pages)

Adrienne Felt will present the paper in Glasgow in April 1.

Abstract
Social networks are increasingly supporting external content integration with platforms such as OpenSocial and the Facebook API. These platforms let users embed third-party applications in their profiles and are a popular example of a mashup. Content integration is often accomplished by proxying the third-party content or importing third-party scripts. However, these methods introduce serious risks of user impersonation and data exposure. Modern browsers provide no mechanism to differentiate between trusted and untrusted embedded content. As a result, content providers are forced to trust third-party scripts or ensure user safety by means of server-side code sanitization. We demonstrate the difficulties of server-side code filtering — and the ramifications of its failure — with an example from the Facebook Platform. We then propose browser modifications that would distinguish between trusted and untrusted content and enforce their separation.

Full Paper

Our paper on using data diversity in a redundant execution framework is now available:

• Anh Nguyen-Tuong, David Evans, John C. Knight, Benjamin Cox, Jack W. Davidson. Security through Redundant Data Diversity. 38^th IEEE/IFPF International Conference on Dependable Systems and Networks, Anchorage, Alaska, June 2008. (PDF, 10 pages; Abstract)

Anh Nguyen-Tuong will present the paper in Anchorage at the DSN conference in June.

This paper builds on our pervious work on N-Variant Systems. It includes the earliest systems-related technical related work in any paper I’ve been involved in, and I doubt we’ll be able to get an earlier one anytime soon. It turns out efforts Nevil Maskelyne led back in the 1700s to compute astronomical tables is closely related to the techniques we describe in this paper.