This repository puts forward a proposal for a secure, decentralized, privacy-preserving presence tracing system. Our proposal aims to simplify and accelerate the process of notifying individuals that shared a semi-public location with a SARS-CoV-2-positive person for a prolonged time without introducing new risks for users and locations. Existing proximity tracing systems (apps for contact tracing such as SwissCovid, Corona Warn App, and Immuni) notify only a subset of these people: those that were close enough for long enough. Current events have shown the need to notify all people that shared a space with a SARS-CoV-2-positive person. The proposed system aims to provide an alternative to increasing use of apps with similar intentions based on invasive collection or that are prone to abuse by authorities. Our design aims to minimize privacy and security risks for individuals and communities, while guaranteeing the highest level of data protection and good usability and deployability.
We continue to seek feedback from a broad audience on the high-level design, its security and privacy properties, and the functionality it offers, so that the merits of the design can be discussed, and protection mechanisms can be added to correct weaknesses. We feel it is essential that designs be made public so the community as a whole can discuss the proposal and verify the claimed privacy guarantees before applications are deployed.
We have recently proposed two new variants of CrowdNotifier:
-
A managed version of CrowdNotifier that enables an organization to manage many locations (for example, meeting rooms) at the same time without the overhead of storing different tracing QR codes for each of them.
-
A server-based version of CrowdNotifier that enables automatically triggering of notifications based on uploads by users. The health authority does not have to be in the loop, increasing speed of notification. At the same time, abuse resistance is weaker than in the full CrowdNotifier protocol.
Both of these versions are compatible with the full CrowdNotifier protocol. Detailed specifications and analysis can be found in the technical specification.
The current version of the white-paper contains several improvements with respect to version v1.0 that we published here earlier. The cryptographic protocol has been changed significantly to avoid the need to reprint a new QR code after triggering a notification. The white paper itself contains a more structured analysis of the security and privacy properties, as well as a more extensive comparison with other deployed systems systems.
An academic paper describing the basic CrowdNotifier scheme was accepted for publication in PoPETs. It contains formal security and privacy definitions and cryptographic proofs that the proposed scheme satisfies them. It also has several smaller improvements over the white paper. Until the camera ready paper is public you can find a preprint version of the paper in this repository.
Please cite this paper as:
@article{LueksGVBSPT21,
author = {Wouter Lueks and
Seda Gürses and
Michael Veale and
Edouard Bugnion and
Marcel Salathé and
Kenneth G. Paterson and
Carmela Troncoso},
title = {CrowdNotifier: Decentralized Privacy-Preserving Presence Tracing},
journal = {Proc. Priv. Enhancing Technol.},
volume = {2021},
number = {4},
year = {2021},
notes = {To appear},
}The technical specification of all three CrowdNotifier protocols can be found here. Source files for this specification are available on our GitHub repository. The source repository is only for reference, please continue using this main documents repository to provide high-level feedback on CrowdNotifier.
The reference implementation of CrowdNotifier implements the cryptographic schemes as described in the white paper. The proof-of-concept SDKs for Android and iOS implement the core cryptographic operations
We are a international consortium of technologists, legal experts, engineers and epidemiologists with a wide range of experience who are interested in ensuring that any presence tracing technology does not result in governments obtaining surveillance capabilities which will endanger civil society.
The following people are behind this design:
EPFL: Prof. Carmela Troncoso, Prof. Marcel Salathé, Prof. Edouard Bugnion, Dr. Wouter Lueks
ETH: Prof. Kenneth G. Paterson
TU Delft: Prof. Seda Gürses
University College London: Dr. Michael Veale
This consortium overlaps with the DP3T consortium, but we stress that these are separate projects.
Everyone interacting on the CrowdNotifier projects codebases, issue trackers, etc. is expected to follow the code of conduct.
We are grateful for the feedback, questions, suggestions and fixes of everyone that commented on earlier versions of this white paper: Paul-Olivier Dehaye, Daniel Kahn Gillmor, Teo Goddet, Andreas Greulich, Matthias Payer, Thusid, Christian Paul, Trond Arve Wasskog, and Peter Wells.