Availability-based methods for distributed storage systems

Anne-Marie Kermarrec; Erwan Le Merrer; Gilles Straub; Alexandre Van Kempen

hal-00521034, version 2

Availability-based methods for distributed storage systems

Anne-Marie Kermarrec () ¹, Erwan Le Merrer () ², Gilles Straub () ², Alexandre Van Kempen () ²

SRDS 2012, 31st International Symposium on Reliable Distributed Systems (2012) 151-160

1: ASAP (INRIA - IRISA)
http://www.inria.fr/equipes/asap
CNRS : UMR6074 – INRIA – Institut National des Sciences Appliquées (INSA) - Rennes – Université de Rennes 1 Campus de Beaulieu 35042 Rennes cedex France
2: Technicolor R & I
Technicolor 1 avenue de Belle Fontaine 35 576 Cesson Sévigné Cedex France France

Available versions : v1 (2010-09-27) v2 (2011-03-07)

Bibliographic reference

Attached file list to this document:

PDF

Availability.pdf

(231.4 KB)

hal-00521034, version 2
http://hal.archives-ouvertes.fr/hal-00521034
oai:hal.archives-ouvertes.fr:hal-00521034
From: Erwan Le Merrer <>
Submitted on: Monday, 7 March 2011 11:33:52
Updated on: Tuesday, 9 October 2012 19:55:38

See short view

Associated documents

PDF :

DOI: 10.1109/SRDS.2012.10

Export

Bibtex EndNote TEI RefWorks

%% hal-00521034, version 2
%% http://hal.archives-ouvertes.fr/hal-00521034
@inproceedings{kermarrec:hal-00521034,
    hal_id = {hal-00521034},
    url = {http://hal.archives-ouvertes.fr/hal-00521034},
    title = {{Availability-based methods for distributed storage systems}},
    author = {Kermarrec, Anne-Marie and Le Merrer, Erwan and Straub, Gilles and Van Kempen, Alexandre},
    abstract = {{Distributed storage systems rely heavily on redundancy to ensure data availability as well as durability. In networked systems subject to intermittent node unavailability, the level of redundancy introduced in the system should be minimized and maintained upon failures. Repairs are well- known to be extremely bandwidth-consuming and it has been shown that, without care, they may significantly congest the system. In this paper, we propose an approach to redundancy management accounting for nodes heterogeneity with respect to availability. We show that by using the availability history of nodes, the performance of two important faces of distributed storage (replica placement and repair) can be significantly improved. Replica placement is achieved based on complementary nodes with respect to nodes availability, improving the overall data availability. Repairs can be scheduled thanks to an adaptive per-node timeout according to node availability, so as to decrease the number of repairs while reaching comparable availability. We propose practical heuristics for those two issues. We evaluate our approach through extensive simulations based on real and well-known availability traces. Results clearly show the benefits of our approach with regards to the critical trade-off between data availability, load-balancing and bandwidth consumption.}},
    keywords = {Distributed storage systems, Availability, timeout},
    language = {English},
    affiliation = {ASAP - INRIA - IRISA , Technicolor R \& I},
    booktitle = {{SRDS 2012, 31st International Symposium on Reliable Distributed Systems}},
    pages = {151-160},
    address = {Irvine, California., United States},
    journal = {International Symposium on Reliable Distributed Systems},
    audience = {international },
    doi = {10.1109/SRDS.2012.10 },
    year = {2012},
    month = Oct,
    pdf = {http://hal.archives-ouvertes.fr/hal-00521034/PDF/Availability.pdf},
}

%F hal-00521034, version 2
%0 Conference Proceedings
%U http://hal.archives-ouvertes.fr/hal-00521034
%2 INFO:INFO_DC;INFO:INFO_DS
%T Availability-based methods for distributed storage systems
%A Kermarrec, Anne-Marie
%A Le Merrer, Erwan
%A Straub, Gilles
%A Van Kempen, Alexandre
%+ ASAP - INRIA - IRISA , Technicolor R & I
%X Distributed storage systems rely heavily on redundancy to ensure data availability as well as durability. In networked systems subject to intermittent node unavailability, the level of redundancy introduced in the system should be minimized and maintained upon failures. Repairs are well- known to be extremely bandwidth-consuming and it has been shown that, without care, they may significantly congest the system. In this paper, we propose an approach to redundancy management accounting for nodes heterogeneity with respect to availability. We show that by using the availability history of nodes, the performance of two important faces of distributed storage (replica placement and repair) can be significantly improved. Replica placement is achieved based on complementary nodes with respect to nodes availability, improving the overall data availability. Repairs can be scheduled thanks to an adaptive per-node timeout according to node availability, so as to decrease the number of repairs while reaching comparable availability. We propose practical heuristics for those two issues. We evaluate our approach through extensive simulations based on real and well-known availability traces. Results clearly show the benefits of our approach with regards to the critical trade-off between data availability, load-balancing and bandwidth consumption.
%G English
%2 2010-09-23
%R 10.1109/SRDS.2012.10
%J International Symposium on Reliable Distributed Systems
%B SRDS 2012, 31st International Symposium on Reliable Distributed Systems
%2 international
%8 2012-10-08
%P 151-160
%B SRDS 2012, 31st International Symposium on Reliable Distributed Systems
%8 2012-10-08
%8 2012-10-11
%2 Irvine, California.
%2 United States
%K Distributed storage systems, Availability, timeout

<?xml version="1.0" encoding="UTF-8"?>
<listBibl xmlns="http://www.tei-c.org/ns/1.0" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mml="http://www.w3.org/1998/Math/MathML">
<listBibl type="inproceedings"><head>Peer-reviewed conferences/proceedings</head>
<biblStruct type="inproceedings" xml:id="hal-00521034"><analytic>
<author><persName><forename>Anne-Marie</forename><surname>Kermarrec</surname></persName>
<affiliation>
<orgName type="laboratory">INRIA - IRISA</orgName>
<orgName type="team">ASAP</orgName>
<country>FR</country>
<orgName type="institution">CNRS : UMR6074</orgName>
<orgName type="institution">INRIA</orgName>
<orgName type="institution">Institut National des Sciences Appliquées (INSA) - Rennes</orgName>
<orgName type="institution">Université de Rennes 1</orgName>
</affiliation>
</author>
<author><persName><forename>Erwan</forename><surname>Le Merrer</surname></persName>
<affiliation>
<orgName type="team">Technicolor R & I</orgName>
<country>FR</country>
<orgName type="institution">Technicolor</orgName>
</affiliation>
</author>
<author><persName><forename>Gilles</forename><surname>Straub</surname></persName>
<affiliation>
<orgName type="team">Technicolor R & I</orgName>
<country>FR</country>
<orgName type="institution">Technicolor</orgName>
</affiliation>
</author>
<author><persName><forename>Alexandre</forename><surname>Van Kempen</surname></persName>
<affiliation>
<orgName type="team">Technicolor R & I</orgName>
<country>FR</country>
<orgName type="institution">Technicolor</orgName>
</affiliation>
</author>
<title type="main" level="a">Availability-based methods for distributed storage systems</title></analytic><monogr><title level="m" type="main">SRDS 2012, 31st International Symposium on Reliable Distributed Systems</title><meeting>SRDS 2012, 31st International Symposium on Reliable Distributed Systems</meeting><title level="j" type="main">International Symposium on Reliable Distributed Systems</title><imprint><pubPlace><settlement>Irvine, California.</settlement><country>United States</country></pubPlace><biblScope type="publicationDate">2012-10-08</biblScope><biblScope type="pp">151-160</biblScope></imprint></monogr><idno type="doi">10.1109/SRDS.2012.10</idno><idno type="HALid">http://hal.archives-ouvertes.fr/hal-00521034</idno><idno type="HALFile">http://hal.archives-ouvertes.fr/hal-00521034/PDF/Availability.pdf</idno></biblStruct>
</listBibl>
</listBibl>