Eric Brigham

Date of Award

2019

Document Type

Thesis

Degree Name

Bachelors

Department

Natural Sciences

First Advisor

Lepinski, Matthew

Area of Concentration

Computer Science

Abstract

The following work can be divided into two distinct and non-overlapping parts. The first addresses secure systems and distributed DNS while the second addresses attack vectors and provides extension on the selfish mining problem into a twoplayer and n-player environment. The abstract for each portion is provided over the next two paragraphs, respectively. The costs of operating a secure system are not always shared proportionately to the benefits that participants receive, which often hinders the adoption of security solutions. Blockchains provide us with an economic framework to distribute costs through cryptographic contracts, however, existing work in building secure applications using blockchains does not take advantage of this framework to align costs and benefits for system participants. One such application is the security of the domain name system. Past attempts to secure the DNS using blockchains suffer from a common problem: the costs of security are not shared proportionately to the benefits that participants receive, because benefits are tied to reading the secured content, while costs are tied to writing it into the blockchain. In this work, we outline a new blockchain model for supporting secure systems in which users pay for the system at a rate proportionate to the benefits they receive. This is achieved by charging participants based on the frequency with which their content is securely retrieved and used by others, rather than the frequency with which they post or update content. The resulting model provides a flexible way to control the compensation and transaction rates of miners, and has applications to security in numerous problem domains. The selfish mining problem was initially proposed as a tactic which a miner with sufficient computational power and connectedness could use to gain currency in excess of their proportional computational power [1]. The initial work outlining this tactic was somewhat groundbreaking, establishing a field of study known as subversive mining. That said, this work was limited by the assumption that there was only one selfish miner, while the rest of the network remained honest. This is a rather lofty assumption, as there is little reason to believe that there is one and only one deviant miner at any given time. In this work we investigate the two player case with analysis and simulation to provide bounds of computation in which a player can deviate profitably as well as provide discussion of the results. We additionally investigate and provide limited theoretical results for the n-player case without bounds of computation for profitability.

Download

Share

COinS