Transaction chains: Achieving serializability with low latency in geo-distributed storage systems

Yang Zhang, Russell Power, Siyuan Zhou, Yair Sovran, Marcos K. Aguilera, Jinyang Li

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

Abstract

Currently, users of geo-distributed storage systems face a hard choice between having serializable transactions with high latency, or limited or no transactions with low latency. We show that it is possible to obtain both serializable transactions and low latency, under two conditions. First, transactions are known ahead of time, permitting an a priori static analysis of conflicts. Second, transactions are structured as transaction chains consisting of a sequence of hops, each hop modifying data at one server. To demonstrate this idea, we built Lynx, a geo-distributed storage system that offers transaction chains, secondary indexes, materialized join views, and geo-replication. Lynx uses static analysis to determine if each hop can execute separately while preserving serializability - if so, a client needs wait only for the first hop to complete, which occurs quickly. To evaluate Lynx, we built three applications: an auction service, a Twitter-like microblogging site and a social networking site. These applications successfully use chains to achieve low latency operation and good throughput.

Original languageEnglish (US)
Title of host publicationSOSP 2013 - Proceedings of the 24th ACM Symposium on Operating Systems Principles
Pages276-291
Number of pages16
DOIs
StatePublished - 2013
Event24th ACM Symposium on Operating Systems Principles, SOSP 2013 - Farmington, PA, United States
Duration: Nov 3 2013Nov 6 2013

Other

Other24th ACM Symposium on Operating Systems Principles, SOSP 2013
CountryUnited States
CityFarmington, PA
Period11/3/1311/6/13

Fingerprint

Static analysis
Servers
Throughput

ASJC Scopus subject areas

  • Software

Cite this

Zhang, Y., Power, R., Zhou, S., Sovran, Y., Aguilera, M. K., & Li, J. (2013). Transaction chains: Achieving serializability with low latency in geo-distributed storage systems. In SOSP 2013 - Proceedings of the 24th ACM Symposium on Operating Systems Principles (pp. 276-291) https://doi.org/10.1145/2517349.2522729

Transaction chains : Achieving serializability with low latency in geo-distributed storage systems. / Zhang, Yang; Power, Russell; Zhou, Siyuan; Sovran, Yair; Aguilera, Marcos K.; Li, Jinyang.

SOSP 2013 - Proceedings of the 24th ACM Symposium on Operating Systems Principles. 2013. p. 276-291.

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

Zhang, Y, Power, R, Zhou, S, Sovran, Y, Aguilera, MK & Li, J 2013, Transaction chains: Achieving serializability with low latency in geo-distributed storage systems. in SOSP 2013 - Proceedings of the 24th ACM Symposium on Operating Systems Principles. pp. 276-291, 24th ACM Symposium on Operating Systems Principles, SOSP 2013, Farmington, PA, United States, 11/3/13. https://doi.org/10.1145/2517349.2522729
Zhang Y, Power R, Zhou S, Sovran Y, Aguilera MK, Li J. Transaction chains: Achieving serializability with low latency in geo-distributed storage systems. In SOSP 2013 - Proceedings of the 24th ACM Symposium on Operating Systems Principles. 2013. p. 276-291 https://doi.org/10.1145/2517349.2522729
Zhang, Yang ; Power, Russell ; Zhou, Siyuan ; Sovran, Yair ; Aguilera, Marcos K. ; Li, Jinyang. / Transaction chains : Achieving serializability with low latency in geo-distributed storage systems. SOSP 2013 - Proceedings of the 24th ACM Symposium on Operating Systems Principles. 2013. pp. 276-291
@inproceedings{e2e4c145ffe04d4a9a3d3900ae9d3184,
title = "Transaction chains: Achieving serializability with low latency in geo-distributed storage systems",
abstract = "Currently, users of geo-distributed storage systems face a hard choice between having serializable transactions with high latency, or limited or no transactions with low latency. We show that it is possible to obtain both serializable transactions and low latency, under two conditions. First, transactions are known ahead of time, permitting an a priori static analysis of conflicts. Second, transactions are structured as transaction chains consisting of a sequence of hops, each hop modifying data at one server. To demonstrate this idea, we built Lynx, a geo-distributed storage system that offers transaction chains, secondary indexes, materialized join views, and geo-replication. Lynx uses static analysis to determine if each hop can execute separately while preserving serializability - if so, a client needs wait only for the first hop to complete, which occurs quickly. To evaluate Lynx, we built three applications: an auction service, a Twitter-like microblogging site and a social networking site. These applications successfully use chains to achieve low latency operation and good throughput.",
author = "Yang Zhang and Russell Power and Siyuan Zhou and Yair Sovran and Aguilera, {Marcos K.} and Jinyang Li",
year = "2013",
doi = "10.1145/2517349.2522729",
language = "English (US)",
isbn = "9781450323888",
pages = "276--291",
booktitle = "SOSP 2013 - Proceedings of the 24th ACM Symposium on Operating Systems Principles",

}

TY - GEN

T1 - Transaction chains

T2 - Achieving serializability with low latency in geo-distributed storage systems

AU - Zhang, Yang

AU - Power, Russell

AU - Zhou, Siyuan

AU - Sovran, Yair

AU - Aguilera, Marcos K.

AU - Li, Jinyang

PY - 2013

Y1 - 2013

N2 - Currently, users of geo-distributed storage systems face a hard choice between having serializable transactions with high latency, or limited or no transactions with low latency. We show that it is possible to obtain both serializable transactions and low latency, under two conditions. First, transactions are known ahead of time, permitting an a priori static analysis of conflicts. Second, transactions are structured as transaction chains consisting of a sequence of hops, each hop modifying data at one server. To demonstrate this idea, we built Lynx, a geo-distributed storage system that offers transaction chains, secondary indexes, materialized join views, and geo-replication. Lynx uses static analysis to determine if each hop can execute separately while preserving serializability - if so, a client needs wait only for the first hop to complete, which occurs quickly. To evaluate Lynx, we built three applications: an auction service, a Twitter-like microblogging site and a social networking site. These applications successfully use chains to achieve low latency operation and good throughput.

AB - Currently, users of geo-distributed storage systems face a hard choice between having serializable transactions with high latency, or limited or no transactions with low latency. We show that it is possible to obtain both serializable transactions and low latency, under two conditions. First, transactions are known ahead of time, permitting an a priori static analysis of conflicts. Second, transactions are structured as transaction chains consisting of a sequence of hops, each hop modifying data at one server. To demonstrate this idea, we built Lynx, a geo-distributed storage system that offers transaction chains, secondary indexes, materialized join views, and geo-replication. Lynx uses static analysis to determine if each hop can execute separately while preserving serializability - if so, a client needs wait only for the first hop to complete, which occurs quickly. To evaluate Lynx, we built three applications: an auction service, a Twitter-like microblogging site and a social networking site. These applications successfully use chains to achieve low latency operation and good throughput.

UR - http://www.scopus.com/inward/record.url?scp=84889660968&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84889660968&partnerID=8YFLogxK

U2 - 10.1145/2517349.2522729

DO - 10.1145/2517349.2522729

M3 - Conference contribution

AN - SCOPUS:84889660968

SN - 9781450323888

SP - 276

EP - 291

BT - SOSP 2013 - Proceedings of the 24th ACM Symposium on Operating Systems Principles

ER -