A performance study of big data workloads in cloud datacenters with network variability

Alexandru Uta, Harry Obaseki

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

Public cloud computing platforms are a cost-effective solution for individuals and organizations to deploy various types of workloads, ranging from scientific applications, business-critical workloads, e-governance to big data applications. Co-locating all such different types of workloads in a single datacenter leads not only to performance degradation, but also to large degrees of performance variability, which is the result of virtualization, resource sharing and congestion. Many studies have already assessed and characterized the degree of resource variability in public clouds. However, we are missing a clear picture on how resource variability impacts big data workloads. In this work, we take a step towards characterizing the behavior of big data workloads under network bandwidth variability. Emulating real-world clouds’ bandwidth distribution, we characterize the performance achieved by running real-world big data applications. We find that most big data workloads are slowed down under network variability scenarios, even those that are not network-bound. Moreover, the maximum average slowdown for the cloud setup with highest variability is 1.48 for CPU-bound workloads, and 1.79 for network-bound workloads.

LanguageEnglish
Title of host publicationICPE 2018 - Companion of the 2018 ACM/SPEC International Conference on Performance Engineering
PublisherAssociation for Computing Machinery, Inc
Pages113-118
Number of pages6
Volume2018-January
ISBN (Electronic)9781450356299
ISBN (Print)9781450356299
DOIs
Publication statusPublished - 2 Apr 2018
Event9th ACM/SPEC International Conference on Performance Engineering, ICPE 2018 - Berlin, Germany
Duration: 9 Apr 201813 Apr 2018

Conference

Conference9th ACM/SPEC International Conference on Performance Engineering, ICPE 2018
CountryGermany
CityBerlin
Period9/04/1813/04/18

Fingerprint

Bandwidth
Cloud computing
Program processors
Big data
Degradation
Costs
Industry
Virtualization

Cite this

Uta, A., & Obaseki, H. (2018). A performance study of big data workloads in cloud datacenters with network variability. In ICPE 2018 - Companion of the 2018 ACM/SPEC International Conference on Performance Engineering (Vol. 2018-January, pp. 113-118). Association for Computing Machinery, Inc. https://doi.org/10.1145/3185768.3186299
Uta, Alexandru ; Obaseki, Harry. / A performance study of big data workloads in cloud datacenters with network variability. ICPE 2018 - Companion of the 2018 ACM/SPEC International Conference on Performance Engineering. Vol. 2018-January Association for Computing Machinery, Inc, 2018. pp. 113-118
@inproceedings{2babc4c04a894b6790e4e35a85099a8d,
title = "A performance study of big data workloads in cloud datacenters with network variability",
abstract = "Public cloud computing platforms are a cost-effective solution for individuals and organizations to deploy various types of workloads, ranging from scientific applications, business-critical workloads, e-governance to big data applications. Co-locating all such different types of workloads in a single datacenter leads not only to performance degradation, but also to large degrees of performance variability, which is the result of virtualization, resource sharing and congestion. Many studies have already assessed and characterized the degree of resource variability in public clouds. However, we are missing a clear picture on how resource variability impacts big data workloads. In this work, we take a step towards characterizing the behavior of big data workloads under network bandwidth variability. Emulating real-world clouds’ bandwidth distribution, we characterize the performance achieved by running real-world big data applications. We find that most big data workloads are slowed down under network variability scenarios, even those that are not network-bound. Moreover, the maximum average slowdown for the cloud setup with highest variability is 1.48 for CPU-bound workloads, and 1.79 for network-bound workloads.",
author = "Alexandru Uta and Harry Obaseki",
year = "2018",
month = "4",
day = "2",
doi = "10.1145/3185768.3186299",
language = "English",
isbn = "9781450356299",
volume = "2018-January",
pages = "113--118",
booktitle = "ICPE 2018 - Companion of the 2018 ACM/SPEC International Conference on Performance Engineering",
publisher = "Association for Computing Machinery, Inc",

}

Uta, A & Obaseki, H 2018, A performance study of big data workloads in cloud datacenters with network variability. in ICPE 2018 - Companion of the 2018 ACM/SPEC International Conference on Performance Engineering. vol. 2018-January, Association for Computing Machinery, Inc, pp. 113-118, 9th ACM/SPEC International Conference on Performance Engineering, ICPE 2018, Berlin, Germany, 9/04/18. https://doi.org/10.1145/3185768.3186299

A performance study of big data workloads in cloud datacenters with network variability. / Uta, Alexandru; Obaseki, Harry.

ICPE 2018 - Companion of the 2018 ACM/SPEC International Conference on Performance Engineering. Vol. 2018-January Association for Computing Machinery, Inc, 2018. p. 113-118.

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - A performance study of big data workloads in cloud datacenters with network variability

AU - Uta, Alexandru

AU - Obaseki, Harry

PY - 2018/4/2

Y1 - 2018/4/2

N2 - Public cloud computing platforms are a cost-effective solution for individuals and organizations to deploy various types of workloads, ranging from scientific applications, business-critical workloads, e-governance to big data applications. Co-locating all such different types of workloads in a single datacenter leads not only to performance degradation, but also to large degrees of performance variability, which is the result of virtualization, resource sharing and congestion. Many studies have already assessed and characterized the degree of resource variability in public clouds. However, we are missing a clear picture on how resource variability impacts big data workloads. In this work, we take a step towards characterizing the behavior of big data workloads under network bandwidth variability. Emulating real-world clouds’ bandwidth distribution, we characterize the performance achieved by running real-world big data applications. We find that most big data workloads are slowed down under network variability scenarios, even those that are not network-bound. Moreover, the maximum average slowdown for the cloud setup with highest variability is 1.48 for CPU-bound workloads, and 1.79 for network-bound workloads.

AB - Public cloud computing platforms are a cost-effective solution for individuals and organizations to deploy various types of workloads, ranging from scientific applications, business-critical workloads, e-governance to big data applications. Co-locating all such different types of workloads in a single datacenter leads not only to performance degradation, but also to large degrees of performance variability, which is the result of virtualization, resource sharing and congestion. Many studies have already assessed and characterized the degree of resource variability in public clouds. However, we are missing a clear picture on how resource variability impacts big data workloads. In this work, we take a step towards characterizing the behavior of big data workloads under network bandwidth variability. Emulating real-world clouds’ bandwidth distribution, we characterize the performance achieved by running real-world big data applications. We find that most big data workloads are slowed down under network variability scenarios, even those that are not network-bound. Moreover, the maximum average slowdown for the cloud setup with highest variability is 1.48 for CPU-bound workloads, and 1.79 for network-bound workloads.

UR - http://www.scopus.com/inward/record.url?scp=85050473611&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050473611&partnerID=8YFLogxK

U2 - 10.1145/3185768.3186299

DO - 10.1145/3185768.3186299

M3 - Conference contribution

SN - 9781450356299

VL - 2018-January

SP - 113

EP - 118

BT - ICPE 2018 - Companion of the 2018 ACM/SPEC International Conference on Performance Engineering

PB - Association for Computing Machinery, Inc

ER -

Uta A, Obaseki H. A performance study of big data workloads in cloud datacenters with network variability. In ICPE 2018 - Companion of the 2018 ACM/SPEC International Conference on Performance Engineering. Vol. 2018-January. Association for Computing Machinery, Inc. 2018. p. 113-118 https://doi.org/10.1145/3185768.3186299