On distributed ledger technologies: designing decentralised and fair algorithmic applications

Abstract

This work focuses on the study of distributed ledger applications, presenting proposals of fair and decentralised applications to counter scenarios in which centralisation of wealth and power are the norm. The first contribution is a novel architecture for a decentralised data market, in which participants crowd-source data and receive a fair share of the reward. The market is shown to be resilient against a number of adversarial behaviours. Subsequently, an algorithm to prove one's location is presented. This algorithm is a key component necessary to the functioning of the data market. In contrast to prior approaches, the design does not require assumptions of honest participation, nor dependence on an external ground truth to identify malicious actors. It is fully peer-to-peer, robust in highly adversarial settings, and compatible with privacy-preserving techniques. The security and reliability of the algorithm are evaluated empirically and characterised mathematically. The protocol is then generalised into a consensus mechanism applicable beyond location verification. An extended mathematical model is developed for this case, and its performance under varying operational conditions is systematically characterised. Finally, a study of governance vulnerabilities in Distributed Ledger Technologies is presented. This work provides a taxonomy of formalised properties necessary for good governance, solutions to implement them and an evaluation of how the absence of these cause severe vulnerabilities. The analysis is then extended to realm of Decentralised Autonomous Organisations (DAOs), which are a class of applications implemented on Distributed Ledger Technologies. The findings anticipated several governance exploits that later materialised, incurring losses in the scale of millions for multiple DAOs. Overall, this thesis aims to contribute to the technological development of distributed ledger applications with the goal of furthering social good, presenting architectures, algorithms, and governance properties that prioritise fairness, decentralisation, and resilience.