Year: 2025
Type: dissertation
Abstract: With their seminal whitepaper "Bitcoin: A Peer-to-Peer Electronic Cash System" and its implementation the author known by the pseudonym Satoshi Nakamoto sparked a chain of events that has led to massive interest and investment in the "blockchain".The stated ambition was to provide an electronic payment system based on cryptographic proofs instead of relying on financial institutions serving as trusted third parties.Apart from removing reliance on centralized authorities the aim was to reduce transaction fees by avoiding costs associated with mediating disputes.If this ambition is to be taken at face value, then it has at least not yet been fulfilled.While the Bitcoin protocol is running a distributed system in a permissionless manner using proofof-work, there are fundamental technical obstacles to realizing the ambition of a decentralized payment system that could be a realistic alternative to services run by financial institutions.We discuss some of the technical challenges and proposed solutions to implementing robust distributed systems in manner that can simultaneously scale to be run by a massive number of parties, operate with sufficiently low latency to be used for everyday transactions, and keep the communication and computational costs at a level that does not result in prohibitive transaction fees.As potential steps towards solutions we propose a series of protocols with asynchronous security.Such protocols assume nothing about the network beyond eventual delivery of messages and simultaneously enjoy the maximal robustness and the lowest possible latency in any given network.They are however associated with significantly higher communication costs than for instance Nakamoto consensus.Our first results are solutions to scalability issues of asynchronous consensus in the context of proofof-stake protocols.These are concretely efficient, but assume static corruption.We also describe a solution with security against adaptive corruption but using heavier cryptographic tools.We combine these into a concretely efficient construction against delayed adaptive corruptions.Finally, we describe a concretely efficient instantiation of a Monte Carlo flavored consensus protocol.As an alternative to the asynchronous solutions we describe a synchronous solution that tolerates more corruption than is possible in asynchrony, but enjoys asynchronous security and responsiveness when the number of corruptions is low.i Resumé Med deres banebrydende whitepaper "Bitcoin: A Peer-to-Peer Electronic Cash System" og implementeringen heraf, satte forfatteren kendt under pseudonymet Satoshi Nakamoto gang i en kæde af begivenheder, der har ført til massiv interesse og investering i "blockchain".Den erklærede ambition var at realisere et elektronisk betalingssystem baseret på kryptografiske beviser i stedet for at stole på finansielle institutioner, der fungerer som betroede tredjeparter.Udover at fjerne behovet for tillid til centrale autoriteter, var målet at reducere transaktionsgebyrer ved at undgå udgifter forbundet med mægling af tvister.Hvis denne ambition tages for pålydende, er den dog endnu ikke blevet opfyldt.Selvom Bitcoin-protokollen kører et distribueret system på et åbent netværk ved brug af proof-of-work, er der grundlæggende tekniske hindringer for at realisere ambitionen om et decentraliseret betalingssystem, der realistisk set kunne være et alternativ til de tjenester, der drives af finansielle institutioner.Vi diskuterer nogle af de tekniske udfordringer og foreslåede løsninger til at implementere robuste distribuerede systemer på en måde, der samtidig kan: skaleres til at blive drevet af et stort antal parter, operere med tilstrækkelig lav forsinkelse til at kunne bruges til hverdagstransaktioner, og holde kommunikations-og beregningsomkostningerne på et niveau, hvor de ikke forårsager for høje transaktionsgebyrer.Som potentielle skridt mod løsninger foreslår vi en række protokoller, der er asynkront sikre.Sådanne protokoller antager intet om netværket udover eventuel levering af beskeder og har samtidig den maksimale robusthed og den lavest mulige forsinkelse i ethvert givet netværk.De er dog forbundet med betydeligt højere kommunikationsomkostninger end for eksempel Nakamoto-konsensus.Vores første resultater er løsninger på skaleringsproblemer ved asynkron konsensus i konteksten af proof-of-stake-protokoller.Disse er konkret effektive, men kun sikre under statisk korruption.Vi beskriver også en løsning med sikkerhed mod adaptiv korruption, men ved brug af tungere kryptografiske værktøjer.Vi kombinerer disse i en konkret effektiv konstruktion mod forsinket adaptiv korruption.Endelig beskriver vi en konkret effektiv implementering af en Monte Carlo-randomiseret konsensusprotokol.Som et alternativ til de asynkrone løsninger beskriver vi en synkron løsning, der tolererer mere korruption, end hvad der er muligt i asynkrone systemer, men har asynkron sikkerhed og responsivitet, når antallet af korruptioner er lavt.iii I have been extraordinarily privileged to spend eight years at the Department of Computer Science, from undergrad to PhD.I want to express my thanks to Simon, Jonas, and Mads who (mostly) stuck around for the first five.To my family and friends, who were always around when I needed them.To the Crypto Group, for existing just next to me, and for being such an incredible group of people I could stumble into.To Sabine, for making me aware of this.To Ivan, for fostering this environment, for introducing me to cryptography, and for his patience, empathy, and humor in teaching and all things.To Malene, for solving every problem with a smile.To Christian Cachin for hosting me, and to all those who made my time in Bern fun and memorable.To cryptographers around the world, whom I had the joy of meeting.
Author: Kamp, Simon Holmgaard
Institution: Aarhus University
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4412211435