Thesis title: Building Private and Secure Exchange Protocols for Decentralized Applications.

Abstract: Adaptor Signatures (AS) extend traditional digital signatures by enabling conditional signing: a pre-signature associated with an instance of a hard relation can be completed into a full signature only upon disclosure of a corresponding witness. This property enables privacy-preserving transaction protocols without requiring changes to the underlying blockchain infrastructure, and serves as a key tool for applications such as atomic swaps and scriptless contracts in blockchains with limited scripting capabilities, such as Bitcoin.

The first contribution of this work is a refined security definition for adaptor signatures that captures the property of witness hiding, ensuring that the signature does not leak information about the witness. To accompany this definition, we present a generic construction of adaptor signatures for any NP relation from one-way functions, thus enhancing both the theoretical foundations and generality of existing schemes.

The second contribution presents an alternative approach for constructing adaptor signature schemes using the Multi-Party Computation in the Head (MPCitH) paradigm. This eliminates the need for costly Karp reductions while preserving a key advantage of MPCitH: the construction treats the underlying cryptographic primitive as a black box (while remaining non-black-box in the relation being proven). Our framework simplifies the design of universal adaptor signatures (UAS) and enhances their applicability across a wide range of decentralized applications, such as blockchain and privacy-preserving systems. Our results demonstrate that MPCitH-based UAS schemes offer strong security guarantees, making them a promising tool in the design of real-world cryptographic protocols.

The third contribution explores the domain of practical secure computation and introduces the MPC Versatile Virtual Machine (mpcVVM), a backend-independent framework and compiler toolchain for automated protocol mixing in multi-party computation. Unlike existing approaches that are manual, heuristic, and tied to a single backend, mpcVVM compiles high-level MPC programs into an intermediate representation, enabling automated, vectorization-aware mixing across multiple backends. We instantiate our framework with MP-SPDZ and MOTION, and validate it against standard benchmarks. Beyond empirical performance gains, we provide a provable optimality guarantee for our mixing methodology and establish tractability under natural structural assumptions on input programs, thereby resolving a problem previously conjectured to be NP-hard. This contribution advances both the systems and theory of MPC, extending provably optimal protocol mixing to the case of multiple protocols.

The fourth contribution of this dissertation extends into the domain of Non-Fungible Tokens (NFTs) and introduces BarterSwap, a fully decentralized protocol for the exchange of multiple NFTs without monetary payments, aiming to establish a barter-style economy within blockchain environments. BarterSwap is built upon the Top Trading Cycles (TTC) algorithm, enabling fair, simultaneous, and mutually beneficial exchanges among multiple participants without the need for intermediaries or a central clearing mechanism. The implementation of the protocol on Ethereum demonstrates its practical viability and shows that it is possible to construct a decentralized model for exchanging digital assets that ensures deterministic correctness, security, and immediate on-chain enforcement of the matching outcome, while operating without any monetary value and relying solely on users’ preferences.

The fifth contribution of this dissertation concerns the design and implementation of a fully decentralized system for Cyber Threat Intelligence (CTI) sharing, which complies with the STIX/TAXII standards and operates on a private Hyperledger Fabric network. The proposed architecture introduces mechanisms for organizational decentralization and controlled access, addressing practical challenges related to integrity, availability, and non-repudiation that characterize centralized TAXII infrastructures.

Supervisor: Aris Pagourtzis

PhD Student: Ioannis Tzannetos