Efficient random network coding for distributed storage systems

Ádám Visegrádi, P. Kacsuk

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Making distributed storage systems reliable is an important challenge. Simple replication may cause severe storage overhead when individual components of The system are very unreliable. Using erasure codes is a promising solution for This problem, but it presents its own challenges; it makes The design of such a system very complex, and it presents The problem of reparation. Network coding has been suggested To be used in The communication in These networks To help reduce overhead. However, using random network coding as -not besides -erasure coding would be an even more promising field To investigate; such a system would have a simple design, need little or no centralization, and reparation of The system could be much simpler Than it is in other erasure coding schemes. The first step on This path is To investigate whether network coding can achieve such a performance That it is a feasible alternative To other erasure codes. This paper presents our experiences about The realization of random network coding based on The discrete logarithm of The finite field. We discuss possible performance optimizations for such a system, and provide performance measurement results focusing on data storage scenarios.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages385-394
Number of pages10
Volume8374 LNCS
ISBN (Print)9783642544194
DOIs
Publication statusPublished - 2014
Event19th International Conference on Parallel Processing Workshops, Euro-Par 2013 - BigDataCloud, DIHC, FedICI, HeteroPar, HiBB, LSDVE, MHPC, OMHI, PADABS, PROPER, Resilience, ROME, and UCHPC 2013 - Aachen, Germany
Duration: Aug 26 2013Aug 27 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8374 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other19th International Conference on Parallel Processing Workshops, Euro-Par 2013 - BigDataCloud, DIHC, FedICI, HeteroPar, HiBB, LSDVE, MHPC, OMHI, PADABS, PROPER, Resilience, ROME, and UCHPC 2013
CountryGermany
CityAachen
Period8/26/138/27/13

Fingerprint

Network coding
Network Coding
Random Networks
Storage System
Distributed Systems
Coding
Discrete Logarithm
Performance Optimization
Performance Measurement
Data Storage
Replication
Galois field
Data storage equipment
Communication
Scenarios
Path
Alternatives

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Visegrádi, Á., & Kacsuk, P. (2014). Efficient random network coding for distributed storage systems. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8374 LNCS, pp. 385-394). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8374 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-642-54420-0_38

Efficient random network coding for distributed storage systems. / Visegrádi, Ádám; Kacsuk, P.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8374 LNCS Springer Verlag, 2014. p. 385-394 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8374 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Visegrádi, Á & Kacsuk, P 2014, Efficient random network coding for distributed storage systems. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 8374 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8374 LNCS, Springer Verlag, pp. 385-394, 19th International Conference on Parallel Processing Workshops, Euro-Par 2013 - BigDataCloud, DIHC, FedICI, HeteroPar, HiBB, LSDVE, MHPC, OMHI, PADABS, PROPER, Resilience, ROME, and UCHPC 2013, Aachen, Germany, 8/26/13. https://doi.org/10.1007/978-3-642-54420-0_38
Visegrádi Á, Kacsuk P. Efficient random network coding for distributed storage systems. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8374 LNCS. Springer Verlag. 2014. p. 385-394. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-54420-0_38
Visegrádi, Ádám ; Kacsuk, P. / Efficient random network coding for distributed storage systems. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8374 LNCS Springer Verlag, 2014. pp. 385-394 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{0f30d6ca958748fd92d5f561e3bf2f6a,
title = "Efficient random network coding for distributed storage systems",
abstract = "Making distributed storage systems reliable is an important challenge. Simple replication may cause severe storage overhead when individual components of The system are very unreliable. Using erasure codes is a promising solution for This problem, but it presents its own challenges; it makes The design of such a system very complex, and it presents The problem of reparation. Network coding has been suggested To be used in The communication in These networks To help reduce overhead. However, using random network coding as -not besides -erasure coding would be an even more promising field To investigate; such a system would have a simple design, need little or no centralization, and reparation of The system could be much simpler Than it is in other erasure coding schemes. The first step on This path is To investigate whether network coding can achieve such a performance That it is a feasible alternative To other erasure codes. This paper presents our experiences about The realization of random network coding based on The discrete logarithm of The finite field. We discuss possible performance optimizations for such a system, and provide performance measurement results focusing on data storage scenarios.",
author = "{\'A}d{\'a}m Visegr{\'a}di and P. Kacsuk",
year = "2014",
doi = "10.1007/978-3-642-54420-0_38",
language = "English",
isbn = "9783642544194",
volume = "8374 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "385--394",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Efficient random network coding for distributed storage systems

AU - Visegrádi, Ádám

AU - Kacsuk, P.

PY - 2014

Y1 - 2014

N2 - Making distributed storage systems reliable is an important challenge. Simple replication may cause severe storage overhead when individual components of The system are very unreliable. Using erasure codes is a promising solution for This problem, but it presents its own challenges; it makes The design of such a system very complex, and it presents The problem of reparation. Network coding has been suggested To be used in The communication in These networks To help reduce overhead. However, using random network coding as -not besides -erasure coding would be an even more promising field To investigate; such a system would have a simple design, need little or no centralization, and reparation of The system could be much simpler Than it is in other erasure coding schemes. The first step on This path is To investigate whether network coding can achieve such a performance That it is a feasible alternative To other erasure codes. This paper presents our experiences about The realization of random network coding based on The discrete logarithm of The finite field. We discuss possible performance optimizations for such a system, and provide performance measurement results focusing on data storage scenarios.

AB - Making distributed storage systems reliable is an important challenge. Simple replication may cause severe storage overhead when individual components of The system are very unreliable. Using erasure codes is a promising solution for This problem, but it presents its own challenges; it makes The design of such a system very complex, and it presents The problem of reparation. Network coding has been suggested To be used in The communication in These networks To help reduce overhead. However, using random network coding as -not besides -erasure coding would be an even more promising field To investigate; such a system would have a simple design, need little or no centralization, and reparation of The system could be much simpler Than it is in other erasure coding schemes. The first step on This path is To investigate whether network coding can achieve such a performance That it is a feasible alternative To other erasure codes. This paper presents our experiences about The realization of random network coding based on The discrete logarithm of The finite field. We discuss possible performance optimizations for such a system, and provide performance measurement results focusing on data storage scenarios.

UR - http://www.scopus.com/inward/record.url?scp=84958537822&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958537822&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-54420-0_38

DO - 10.1007/978-3-642-54420-0_38

M3 - Conference contribution

SN - 9783642544194

VL - 8374 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 385

EP - 394

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -