Efficient verification for provably secure storage and secret sharing in systems where half the servers are faulty

Rida Bazzi, Goran Konjevod

Research output: Contribution to journalArticle

Abstract

We present new decentralized storage systems that are resilient to arbitrary failures of up to a half of all servers and can tolerate a computationally unbounded adversary. These are the first such results with space requirements smaller than those of full replication without relying on cryptographic assumptions. We also significantly reduce share sizes for robust secret-sharing schemes with or without an honest dealer, again without cryptographic assumptions. A major ingredient in our systems is an information verification scheme that replaces hashing (for storage systems) or information checking protocols (for secret sharing). Together with a new way of organizing verification information, this allows us to use a simple majority algorithm to identify with high probability all servers whose information hasn't been corrupted.

Original languageEnglish (US)
Pages (from-to)420-434
Number of pages15
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3274
StatePublished - 2004

Fingerprint

Secret Sharing
Information Systems
Servers
Server
Storage System
Network protocols
Secret Sharing Scheme
Hashing
Decentralized
Replication
Requirements
Arbitrary

Keywords

  • Byzantine failures
  • Secret sharing
  • Secure storage

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

@article{ce436584b5e34cb38461d42b2ee72a04,
title = "Efficient verification for provably secure storage and secret sharing in systems where half the servers are faulty",
abstract = "We present new decentralized storage systems that are resilient to arbitrary failures of up to a half of all servers and can tolerate a computationally unbounded adversary. These are the first such results with space requirements smaller than those of full replication without relying on cryptographic assumptions. We also significantly reduce share sizes for robust secret-sharing schemes with or without an honest dealer, again without cryptographic assumptions. A major ingredient in our systems is an information verification scheme that replaces hashing (for storage systems) or information checking protocols (for secret sharing). Together with a new way of organizing verification information, this allows us to use a simple majority algorithm to identify with high probability all servers whose information hasn't been corrupted.",
keywords = "Byzantine failures, Secret sharing, Secure storage",
author = "Rida Bazzi and Goran Konjevod",
year = "2004",
language = "English (US)",
volume = "3274",
pages = "420--434",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Efficient verification for provably secure storage and secret sharing in systems where half the servers are faulty

AU - Bazzi, Rida

AU - Konjevod, Goran

PY - 2004

Y1 - 2004

N2 - We present new decentralized storage systems that are resilient to arbitrary failures of up to a half of all servers and can tolerate a computationally unbounded adversary. These are the first such results with space requirements smaller than those of full replication without relying on cryptographic assumptions. We also significantly reduce share sizes for robust secret-sharing schemes with or without an honest dealer, again without cryptographic assumptions. A major ingredient in our systems is an information verification scheme that replaces hashing (for storage systems) or information checking protocols (for secret sharing). Together with a new way of organizing verification information, this allows us to use a simple majority algorithm to identify with high probability all servers whose information hasn't been corrupted.

AB - We present new decentralized storage systems that are resilient to arbitrary failures of up to a half of all servers and can tolerate a computationally unbounded adversary. These are the first such results with space requirements smaller than those of full replication without relying on cryptographic assumptions. We also significantly reduce share sizes for robust secret-sharing schemes with or without an honest dealer, again without cryptographic assumptions. A major ingredient in our systems is an information verification scheme that replaces hashing (for storage systems) or information checking protocols (for secret sharing). Together with a new way of organizing verification information, this allows us to use a simple majority algorithm to identify with high probability all servers whose information hasn't been corrupted.

KW - Byzantine failures

KW - Secret sharing

KW - Secure storage

UR - http://www.scopus.com/inward/record.url?scp=35048863566&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35048863566&partnerID=8YFLogxK

M3 - Article

VL - 3274

SP - 420

EP - 434

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -