Files in this item

FilesDescriptionFormat

application/pdf

application/pdfCORDI-THESIS-2017.pdf (2MB)
(no description provided)PDF

Description

Title:Simulating high-throughput cryptocurrency payment channel networks
Author(s):Cordi, Christopher Neal
Advisor(s):Miller, Andrew
Department / Program:Computer Science
Discipline:Computer Science
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):Bitcoin
Ethereum
Blockchains
Abstract:Payment channels secured with cryptocurrency as collateral enable users to make many transactions with few blockchain broadcasts. Networks of payment channels have emerged as a proposed solution to Bitcoin’s scaling problem. Since the proposal of the first payment channel network, the Lightning Network, alternatives promising significant improvements, such as the Sprites protocol, have been proposed. Without at-scale implementations to analyze in situ, it is difficult to make meaningful comparisons of payment channel network protocols. In order to bridge this gap, we introduce a new simulation framework that can be used to evaluate how different payment channel network protocols will perform in both the expected and worst cases. Our framework is generic and accommodates benchmarking across different variants of payment channel network protocols, network topologies, routing algorithms, and user behaviors. User spending behavior in our payment channel network simulator is generated based on behavioral modeling techniques used in credit card fraud research. Our simulation is the first payment channel network simulator to seed user behaviors with data from real-world credit card users. Our framework can be used to evaluate expected case performance and resiliency to attacks across different payment channel network protocols and routing algorithms. We demonstrate the utility of our framework through comparisons of the Lightning Network to Sprites. We also compare the proposed decentralized routing algorithm, Flare, to an ideal centralized routing algorithm. Our results reveal that if spending behaviors are similar to those of credit card users, scale-free network topologies achieve higher throughput and resiliency compared to small-world networks. We also confirm that the Sprites protocol enjoys numerous advantages over the Lightning Network including smaller durations, shorter path length payments, and greater resiliency, all of which are most significant in decentralized topologies using decentralized routing algorithms.
Issue Date:2017-11-16
Type:Text
URI:http://hdl.handle.net/2142/99319
Rights Information:Copyright 2017 Christopher Cordi
Date Available in IDEALS:2018-03-13
Date Deposited:2017-12


This item appears in the following Collection(s)

Item Statistics