Abstract

Blockchain is an evolving technology with the potential to reshape various industries. It is an immutable append-only distributed ledger that maintains the integrity and availability of its transactions. With blockchain, mutually distrusting parties can finally make transactions without relying on a trusted third party. Nevertheless, many organizations are reluctant to adopt it due to several issues such as privacy. More precisely, the inherent transparency of transactions in blockchain comes at the cost of privacy despite the use of pseudonymous identities. We design cryptographic protocols to improve the privacy of a set of decentralized applications utilizing blockchain.

The rapidly growing number of digital assets deployed over blockchain requires a convenient trading mechanism. Sealed-bid auctions are powerful trading tools due to their privacy advantages compared to their open-cry counterparts. However, the inherent transparency on the blockchain makes designing a sealed-bid auction a challenging task. We propose three protocols utilizing zero-knowledge proofs, trusted execution environments, and smart contracts to publicly verify the correctness of the auction winner while maintaining users' privacy.

In the first protocol, the auctioneer utilizes zero-knowledge proof of interval membership to prove the correctness of the auction winner without revealing the losing bids. However, this protocol is expensive in verification cost and scales linearly with the number of users. To reduce the verification cost, we design a second protocol where the auctioneer utilizes an advanced zero-knowledge proving system with a constant verification complexity. Both protocols offer partial privacy as the auctioneer gets to know the actual values of bids. The third protocol provides complete privacy by utilizing a trusted execution environment to determine the auction winner without revealing the losing bids to any party. Furthermore, since this protocol relies on simple cryptographic primitives, it achieves the lowest verification cost with a constant complexity regardless of the number of bids.

Extending the work on sealed-bid auctions, we tackle a privacy problem in lit markets where all the information about bids and offers in the order book is visible to the public. While transparency helps the price discovery, it hurts financial institutions that trade large bulk orders. Therefore, we design a privacy-preserving periodic auction that hides limit-orders during the submission phase while preventing front-running and ensuring the correctness of market-clearing prices.

Next, we target a privacy problem in inter-bank payment systems. Banks transfer money and securities instantaneously on a gross basis by utilizing Real-Time Gross Settlement (RTGS) system. Central banks operate RTGS systems and require access to payment instructions of each local inter-bank. Accordingly, RTGS systems assume unconditional trust given to central banks, and they suffer from a single point of failure. Hence, we propose a decentralized netting protocol that ensures balance correctness while hiding the transferred amounts and recipients.

Finally, we switch gears to the booming Non-Fungible Tokens (NFTs) technology and tackle privacy issues with existing systems. NFTs are unique non-interchangeable digital assets verified and secured by blockchain technology. Current NFT standards lack privacy guarantees; hence any observer can trivially learn the whole NFT collection of an arbitrary user. Furthermore, popular marketplaces use public exchanges and auctions for trades which leak information about the trade parties and the payment amount for an NFT. We design Aegis as a protocol that adds privacy to NFTs ownership. More importantly, Aegis allows users to atomically swap NFTs for payment amounts while hiding the details of the transactions.