Files in this item

FilesDescriptionFormat

application/pdf

application/pdfRANA-THESIS-2020.pdf (418kB)
(no description provided)PDF

Description

Title:Barracuda: The power of l-polling in proof of stake blockchains
Author(s):Rana, Ranvir
Advisor(s):Viswanath, Pramod
Department / Program:Electrical & Computer Eng
Discipline:Electrical & Computer Engr
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):Blockchain
Distributed, Parallel, and Cluster Computing
Cryptography and Security
Information Theory
Abstract:Blockchain is a database of storing sequential events as a chain of blocks consistently across a distributed set of nodes. A fundamental problem in doing so is to decide where to put the next block and who should do it in a Sybil-resistant manner. To solve this problem, typically, a node is elected randomly as a leader to append a new block to the end of a chain stored locally by the leader. Ideally, this should extend the chain of blocks, however in practice, due to network imperfections, the local blockchain of the leader might not be synced entirely, thus resulting in forking, a scenario when a new block is appended in the middle of the blockchain, thus creating a fork. These network imperfections create a structure like a tree rather than a chain, where blocks not part of the main chain are abandoned, thus reducing the system’s efficiency. We propose a new peer-to-peer (P2P) protocol called Barracuda, where the leader polls l − 1 random nodes for their blocktree information before proposing a new block and show that this policy has an effect equivalent to having a network that is l times faster under a stochastic network model inspired by Decker and Wattenhofer (2013). We also show via simulations that Barracuda is robust to several real-world factors in the network model.
Issue Date:2020-11-30
Type:Thesis
URI:http://hdl.handle.net/2142/109389
Rights Information:Copyright 2020 Ranvir Rana
Date Available in IDEALS:2021-03-05
Date Deposited:2020-12


This item appears in the following Collection(s)

Item Statistics