This project focused on the problem of analyzing, formalizing, and enhancing privacy in blockchain systems and applications. 

Intellectual merits.

 Our main contributions can be grouped as follows:

• Analyzing privacy: We  proved a lower bound on the highest anonymity guarantees that can be achieved in any proof-of-stake blockchain. We showed that de-anonymization attacks can be mounted through network delays, regardless of whether parties are using anonymous broadcast or point-to-point channels, and regardless of the cryptographic protocols being employed. As a consequence we proved that  full anonymity is impossible to achieve in network that allow adversarial delays (S&P ‘21). Along the same lines, but in a different contest, we showed other privacy attacks that are based on the ability of the adversary to run smart contracts, and allow an adversary  to break the zero-knowledge property of certain blockchain-based zero-knowledge proofs (PKC ‘21,’20).

• Formalizing privacy:  we explored novel approaches for defining anonymity for blockchain-based payment systems, based on differential privacy (PoPETS ‘22). We provided the first definition of privacy for side-chains (CBT workshop at ESORICS ’21), and provided a framework to categorize and analyze techniques for accountability and auditability of anonymous blockchains (ACNS ’21).

• Enhancing  privacy: We introduced and constructed new building blocks to enhance privacy in blockchain applications. The most recent primitive,  called Private Signaling, was awarded the distinguished paper award at USENIX ‘22 and it aims at providing  full anonymity with zero-overhead to blockchain users. We also built a plethora of primitives that serve as building blocks for adding privacy under various trade-offs in terms of efficiency, trust assumptions and computational assumptions. In  terms of efficiency, we constructed communication-efficient cryptographic accumulators in the Bilinear Pairing setting, that allows for faster batching and aggregation of zero-knowledge proofs (CCS ‘22). Additionally, we build an efficient system for proving total assets in cryptocurrencies in a privacy preserving way (PoPETS ‘22). In terms of reducing trust-assumptions, we constructed publicly verifiable zero-knowledge proofs that do not require any trusted setup, but only the existence of a blockchain with some unpredictability property. We also designed a decentralized protocol that allows a crowd of people to audit the validity of a trusted setup process even in the setting that all servers and all the clients of the MPC protocol are subverted by an adversary (ASIACRYPT ‘20). In terms of reducing computational assumptions, we built the first post-quantum secure threshold ring signature (PKC ‘21), and one-time traceable ring signatures  that use primitives in a black-box manner (ESORICS21). Notably, the latter is the only example of an anonymous building block that uses a random oracle only.

Broader Impacts

Our results have direct impacts on both the blockchain research and development communities. We provide analysis on the inherent limitations of anonymity (e.g., through our lower bound) and what a variety of tools that can be used to enhance anonymity in a provably secure way (e.g., through our building blocks). Importantly, all our results are proved in formal frameworks.

The PIs have actively engaged with local blockchain interest groups, startup and established companies working on blockchain and privacy projects as well as with scientists from different disciplines such as policy and economics researchers.

The results of this proposal have been published in major conferences and are all publicly available to public archives.  We have further disseminated our results, through recorded seminars, tutorials, invited talks and conference presentations.

Last Modified: 02/16/2023
Modified by: Foteini Baldimtsi