Overall Objectives
Research Program
Application Domains
Highlights of the Year
New Software and Platforms
New Results
Bilateral Contracts and Grants with Industry
Partnerships and Cooperations
XML PDF e-pub
PDF e-Pub

Section: New Results

Results for Axis 3: Building a secure network stack

Privacy-Preserving Abuse Detection in Future Decentralised Online Social Networks

Participants : Jeffrey Burdges, Alvaro Garcia Recuero, Christian Grothoff.

Future online social networks need to not only protect sensitive data of their users, but also protect them from abusive behavior coming from malicious participants in the network. We investigated the use of supervised learning techniques to detect abusive behavior and describe privacy-preserving protocols to compute the feature set required by abuse classification algorithms in a secure and privacy-preserving way. While our method is not yet fully resilient against a strong adaptive adversary, our evaluation suggests that it will be useful to detect abusive behavior with a minimal impact on privacy.

Our results show how to combine local knowledge with private set intersection and union cardinality protocols (with masking of BLS signature to protect identity of signers/subscribers) to privately derive feature values from users in OSNs. Given an adaptive adversary that would be able to manipulate most features we propose in our supervised learning approach, it is surprising that with just three features resistant to adversarial manipulation, the algorithms still provide useful classifications.

This work was originally presented at DPM 2016 [63] and expanded upon in Álvaro García-Recuero's PhD thesis [1].

Fog of Trust

Participants : Jeffrey Burdges, Christian Grothoff.

The Web of Trust (WoT) used traditionally used by tools for private communication such as PGP is used to to validate individual links between participants. Using the WoT, however, leaks meta data, such that users must opt-in for it – exposing themselves to risks of privacy loss. We proposed a new method, the Fog of Trust (FoT), which uses the privacy-preserving set intersection cardinality protocol originally used in our work on abuse detection in online social networks, to support this critical step of public key verification via collaboration. In the FoT, the social relationships — which are used to verify public keys – remain hidden. This allows keys to be verified via trusted intermediaries that were established beforehand, without the need to verify each individual new contact using Trustwords. Consequently, FoT will can the same functionality as the WoT without its drawbacks to privacy.

Cell tower privacy

Participants : Christian Grothoff, Neal Walfield.

Context-aware applications are programs that are able to improve their performance by adapting to the current conditions, which include the user's behavior, networking conditions, and charging opportunities. In many cases, the user’s location is an excellent predictor of the context. Thus, by predicting the user’s future location, we can predict the future conditions. In this work, we developed techniques to identify and predict the user's location over the next 24 hours with a minimum median accuracy of 82results include our observation that cell phones sample the towers in their vicinity, which makes cell towers as-is inappropriate for use as landmarks. Motivated by this observation, we developed two techniques for processing the cell tower traces so that landmarks more closely correspond to locations, and cell tower transitions more closely correspond to user movement. We developed a prediction engine, which is based on simple sampling distributions of the form f(t,c), where t is the predicted tower, and c is a set of conditions. The conditions that we considered include the time of the day, the day of the week, the current regime, and the current tower. Our family of algorithms, called TomorrowToday, achieves 89% prediction precision across all prediction trials for predictions 30 minutes in the future. This decreases slowly for predictions further in the future, and levels off for predictions approximately 4 hours in the future, at which point we achieve 82% prediction precision across all prediction trials up to 24 hours in the future. This represents a significant improvement over NextPlace, a well-cited prediction algorithm based on non-linear time series, which achieves appropriately 80% prediction precision (self reported) for predictions 30 minutes in the future, but, unlike our predictors, which try all prediction attempts, NextPlace only attempts 7% of the prediction trials on our data set [67].

Taler protocol improvements

Participants : Jeffrey Burdges, Florian Dold, Christian Grothoff, Marcello Stanisci.

We started modeling the Taler protocol in the framework of Provable Security, precisely defining the formal meaning of income transparency, fairness, anonymity and unforgeablity as security games. The resulting definitions and security proofs allow a more precise statement of the security of Taler in relation to the security assumptions that are being made.

The implementation of the wallet module now supports the full Taler protocol, including the refresh operation for highly efficient and privacy-preserving change.

In addition to improving the stability of the implementation of all Taler components, we added new features to the protocol that (1) allow refunds from merchants without violating privacy and (2) allow merchants to do "customer tipping", which transfers money from merchants directly to customers' wallets as a reward for doing actions on their website.

Mix Networking

Participants : Jeffrey Burdges, Christian Grothoff.

We have begun implementing our ratcheting scheme for providing hybrid post-quantum and forward security to the Sphinx mix network packet format. We also began collaborating with the Panoramix project and LEAP to help resolve numerous practical challenges to deploying a mix network. We shall speak about this ongoing work at the Chaos Computer Club's annual congress 34c3 in December 2017.