OpuS

Fair and efficient cache sharing for in-memory data analytics

Yinghao Yu, Wei Wang, Jun Zhang, Qizhen Weng, Khaled Letaief

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

Abstract

We study the fair cache allocation problem in shared cloud environments, where many users and applications contend for the main memory to cache shared datasets or files. Unlike other resources such as CPUs and networks, in-memory caches can be non-exclusively shared across many users, e.g., a cached columnar dataset queried by many Spark SQL jobs. This results in a unique challenge of the 'free-riding' problem, where a user lies about its caching preferences to trick other users to cache files for it, using their allocated cache space. We show that existing cache allocation policies either suffer from such manipulations or result in poor efficiency. To address this problem, we propose a new cache allocation algorithm, termed OpuS, or Opportunistic Sharing for high efficiency. We show that OpuS provides performance isolation between users and is strategy-proof against 'free-riding' manipulations. We have implemented OpuS as a pluggable cache manager in Alluxio, a popular memory-centric filesystem. Cluster deployment and trace-driven simulations demonstrate that OpuS allocates each user a fair share of caches while achieving near-optimal efficiency in cache utilization.

Original languageEnglish
Title of host publicationProceedings - 2018 IEEE 38th International Conference on Distributed Computing Systems, ICDCS 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages154-164
Number of pages11
Volume2018-July
ISBN (Electronic)9781538668719
DOIs
Publication statusPublished - 19 Jul 2018
Externally publishedYes
Event38th IEEE International Conference on Distributed Computing Systems, ICDCS 2018 - Vienna, Austria
Duration: 2 Jul 20185 Jul 2018

Other

Other38th IEEE International Conference on Distributed Computing Systems, ICDCS 2018
CountryAustria
CityVienna
Period2/7/185/7/18

Fingerprint

Data storage equipment
Cache memory
Electric sparks
Program processors
Managers

Keywords

  • Cache sharing
  • Sharing incentive
  • Strategyproof

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

Yu, Y., Wang, W., Zhang, J., Weng, Q., & Letaief, K. (2018). OpuS: Fair and efficient cache sharing for in-memory data analytics. In Proceedings - 2018 IEEE 38th International Conference on Distributed Computing Systems, ICDCS 2018 (Vol. 2018-July, pp. 154-164). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICDCS.2018.00025

OpuS : Fair and efficient cache sharing for in-memory data analytics. / Yu, Yinghao; Wang, Wei; Zhang, Jun; Weng, Qizhen; Letaief, Khaled.

Proceedings - 2018 IEEE 38th International Conference on Distributed Computing Systems, ICDCS 2018. Vol. 2018-July Institute of Electrical and Electronics Engineers Inc., 2018. p. 154-164.

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

Yu, Y, Wang, W, Zhang, J, Weng, Q & Letaief, K 2018, OpuS: Fair and efficient cache sharing for in-memory data analytics. in Proceedings - 2018 IEEE 38th International Conference on Distributed Computing Systems, ICDCS 2018. vol. 2018-July, Institute of Electrical and Electronics Engineers Inc., pp. 154-164, 38th IEEE International Conference on Distributed Computing Systems, ICDCS 2018, Vienna, Austria, 2/7/18. https://doi.org/10.1109/ICDCS.2018.00025
Yu Y, Wang W, Zhang J, Weng Q, Letaief K. OpuS: Fair and efficient cache sharing for in-memory data analytics. In Proceedings - 2018 IEEE 38th International Conference on Distributed Computing Systems, ICDCS 2018. Vol. 2018-July. Institute of Electrical and Electronics Engineers Inc. 2018. p. 154-164 https://doi.org/10.1109/ICDCS.2018.00025
Yu, Yinghao ; Wang, Wei ; Zhang, Jun ; Weng, Qizhen ; Letaief, Khaled. / OpuS : Fair and efficient cache sharing for in-memory data analytics. Proceedings - 2018 IEEE 38th International Conference on Distributed Computing Systems, ICDCS 2018. Vol. 2018-July Institute of Electrical and Electronics Engineers Inc., 2018. pp. 154-164
@inproceedings{9d97f073b1e14a59be3ae8f6bcdb6b57,
title = "OpuS: Fair and efficient cache sharing for in-memory data analytics",
abstract = "We study the fair cache allocation problem in shared cloud environments, where many users and applications contend for the main memory to cache shared datasets or files. Unlike other resources such as CPUs and networks, in-memory caches can be non-exclusively shared across many users, e.g., a cached columnar dataset queried by many Spark SQL jobs. This results in a unique challenge of the 'free-riding' problem, where a user lies about its caching preferences to trick other users to cache files for it, using their allocated cache space. We show that existing cache allocation policies either suffer from such manipulations or result in poor efficiency. To address this problem, we propose a new cache allocation algorithm, termed OpuS, or Opportunistic Sharing for high efficiency. We show that OpuS provides performance isolation between users and is strategy-proof against 'free-riding' manipulations. We have implemented OpuS as a pluggable cache manager in Alluxio, a popular memory-centric filesystem. Cluster deployment and trace-driven simulations demonstrate that OpuS allocates each user a fair share of caches while achieving near-optimal efficiency in cache utilization.",
keywords = "Cache sharing, Sharing incentive, Strategyproof",
author = "Yinghao Yu and Wei Wang and Jun Zhang and Qizhen Weng and Khaled Letaief",
year = "2018",
month = "7",
day = "19",
doi = "10.1109/ICDCS.2018.00025",
language = "English",
volume = "2018-July",
pages = "154--164",
booktitle = "Proceedings - 2018 IEEE 38th International Conference on Distributed Computing Systems, ICDCS 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - OpuS

T2 - Fair and efficient cache sharing for in-memory data analytics

AU - Yu, Yinghao

AU - Wang, Wei

AU - Zhang, Jun

AU - Weng, Qizhen

AU - Letaief, Khaled

PY - 2018/7/19

Y1 - 2018/7/19

N2 - We study the fair cache allocation problem in shared cloud environments, where many users and applications contend for the main memory to cache shared datasets or files. Unlike other resources such as CPUs and networks, in-memory caches can be non-exclusively shared across many users, e.g., a cached columnar dataset queried by many Spark SQL jobs. This results in a unique challenge of the 'free-riding' problem, where a user lies about its caching preferences to trick other users to cache files for it, using their allocated cache space. We show that existing cache allocation policies either suffer from such manipulations or result in poor efficiency. To address this problem, we propose a new cache allocation algorithm, termed OpuS, or Opportunistic Sharing for high efficiency. We show that OpuS provides performance isolation between users and is strategy-proof against 'free-riding' manipulations. We have implemented OpuS as a pluggable cache manager in Alluxio, a popular memory-centric filesystem. Cluster deployment and trace-driven simulations demonstrate that OpuS allocates each user a fair share of caches while achieving near-optimal efficiency in cache utilization.

AB - We study the fair cache allocation problem in shared cloud environments, where many users and applications contend for the main memory to cache shared datasets or files. Unlike other resources such as CPUs and networks, in-memory caches can be non-exclusively shared across many users, e.g., a cached columnar dataset queried by many Spark SQL jobs. This results in a unique challenge of the 'free-riding' problem, where a user lies about its caching preferences to trick other users to cache files for it, using their allocated cache space. We show that existing cache allocation policies either suffer from such manipulations or result in poor efficiency. To address this problem, we propose a new cache allocation algorithm, termed OpuS, or Opportunistic Sharing for high efficiency. We show that OpuS provides performance isolation between users and is strategy-proof against 'free-riding' manipulations. We have implemented OpuS as a pluggable cache manager in Alluxio, a popular memory-centric filesystem. Cluster deployment and trace-driven simulations demonstrate that OpuS allocates each user a fair share of caches while achieving near-optimal efficiency in cache utilization.

KW - Cache sharing

KW - Sharing incentive

KW - Strategyproof

UR - http://www.scopus.com/inward/record.url?scp=85050988888&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050988888&partnerID=8YFLogxK

U2 - 10.1109/ICDCS.2018.00025

DO - 10.1109/ICDCS.2018.00025

M3 - Conference contribution

VL - 2018-July

SP - 154

EP - 164

BT - Proceedings - 2018 IEEE 38th International Conference on Distributed Computing Systems, ICDCS 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -