A model predictive controller for contention-aware resource allocation in virtualized data centers

M. Reza Hoseinyfarahabady; Young Choon Lee; Albert Y. Zomaya; Zahir Tari; Andy Song

doi:10.1109/MASCOTS.2016.59

A model predictive controller for contention-aware resource allocation in virtualized data centers

M. Reza Hoseinyfarahabady, Young Choon Lee, Albert Y. Zomaya, Zahir Tari, Andy Song

School of Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › peer-review

2 Citations (Scopus)

Abstract

Data center efficiency is primarily sought by sharing physical resources, such as processors, memory, and disks in the form of virtual machines or containers among multiple users, i.e., workload consolidation. However, the reality is co-located applications in these virtual platforms compete for resources and interfere with each others' performance, resulting in performance variability/degradation. In this paper, we present the contentionaware resource allocation (CARA) solution, which optimizes data center efficiency. It is essentially devised based on a model predictive control that enables to make judicious consolidation decisions with future system states. CARA consolidates workloads explicitly taking into account the correlation between shared and isolated resource usage patterns. Based on our experimental results, CARA improves the overall resource utilization by 32%, without a significant impact on the quality-of-service (QoS) enforcement level. Such improvement results in a fewer number of active servers and in turn contributes to an overall energy saving by 33%.

Original language	English
Title of host publication	Proceedings - 2016 IEEE 24th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016
Place of Publication	Piscataway, NJ
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	277-282
Number of pages	6
ISBN (Electronic)	9781509034314
DOIs	https://doi.org/10.1109/MASCOTS.2016.59
Publication status	Published - 5 Dec 2016
Event	24th IEEE International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016 - London, United Kingdom Duration: 19 Sept 2016 → 21 Sept 2016

Other

Other	24th IEEE International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016
Country/Territory	United Kingdom
City	London
Period	19/09/16 → 21/09/16

Access to Document

10.1109/MASCOTS.2016.59

Cite this

Hoseinyfarahabady, M. R., Lee, Y. C., Zomaya, A. Y., Tari, Z., & Song, A. (2016). A model predictive controller for contention-aware resource allocation in virtualized data centers. In Proceedings - 2016 IEEE 24th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016 (pp. 277-282). Article 7774590 Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/MASCOTS.2016.59

Hoseinyfarahabady, M. Reza ; Lee, Young Choon ; Zomaya, Albert Y. et al. / A model predictive controller for contention-aware resource allocation in virtualized data centers. Proceedings - 2016 IEEE 24th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016. Piscataway, NJ : Institute of Electrical and Electronics Engineers (IEEE), 2016. pp. 277-282

@inproceedings{b3530bf62cca4cd28bb119838effd1ad,

title = "A model predictive controller for contention-aware resource allocation in virtualized data centers",

abstract = "Data center efficiency is primarily sought by sharing physical resources, such as processors, memory, and disks in the form of virtual machines or containers among multiple users, i.e., workload consolidation. However, the reality is co-located applications in these virtual platforms compete for resources and interfere with each others' performance, resulting in performance variability/degradation. In this paper, we present the contentionaware resource allocation (CARA) solution, which optimizes data center efficiency. It is essentially devised based on a model predictive control that enables to make judicious consolidation decisions with future system states. CARA consolidates workloads explicitly taking into account the correlation between shared and isolated resource usage patterns. Based on our experimental results, CARA improves the overall resource utilization by 32%, without a significant impact on the quality-of-service (QoS) enforcement level. Such improvement results in a fewer number of active servers and in turn contributes to an overall energy saving by 33%.",

author = "Hoseinyfarahabady, {M. Reza} and Lee, {Young Choon} and Zomaya, {Albert Y.} and Zahir Tari and Andy Song",

year = "2016",

month = dec,

day = "5",

doi = "10.1109/MASCOTS.2016.59",

language = "English",

pages = "277--282",

booktitle = "Proceedings - 2016 IEEE 24th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

address = "United States",

note = "24th IEEE International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016 ; Conference date: 19-09-2016 Through 21-09-2016",

}

Hoseinyfarahabady, MR, Lee, YC, Zomaya, AY, Tari, Z & Song, A 2016, A model predictive controller for contention-aware resource allocation in virtualized data centers. in Proceedings - 2016 IEEE 24th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016., 7774590, Institute of Electrical and Electronics Engineers (IEEE), Piscataway, NJ, pp. 277-282, 24th IEEE International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016, London, United Kingdom, 19/09/16. https://doi.org/10.1109/MASCOTS.2016.59

A model predictive controller for contention-aware resource allocation in virtualized data centers. / Hoseinyfarahabady, M. Reza; Lee, Young Choon; Zomaya, Albert Y. et al.
Proceedings - 2016 IEEE 24th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016. Piscataway, NJ: Institute of Electrical and Electronics Engineers (IEEE), 2016. p. 277-282 7774590.

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › peer-review

TY - GEN

T1 - A model predictive controller for contention-aware resource allocation in virtualized data centers

AU - Hoseinyfarahabady, M. Reza

AU - Lee, Young Choon

AU - Zomaya, Albert Y.

AU - Tari, Zahir

AU - Song, Andy

PY - 2016/12/5

Y1 - 2016/12/5

N2 - Data center efficiency is primarily sought by sharing physical resources, such as processors, memory, and disks in the form of virtual machines or containers among multiple users, i.e., workload consolidation. However, the reality is co-located applications in these virtual platforms compete for resources and interfere with each others' performance, resulting in performance variability/degradation. In this paper, we present the contentionaware resource allocation (CARA) solution, which optimizes data center efficiency. It is essentially devised based on a model predictive control that enables to make judicious consolidation decisions with future system states. CARA consolidates workloads explicitly taking into account the correlation between shared and isolated resource usage patterns. Based on our experimental results, CARA improves the overall resource utilization by 32%, without a significant impact on the quality-of-service (QoS) enforcement level. Such improvement results in a fewer number of active servers and in turn contributes to an overall energy saving by 33%.

AB - Data center efficiency is primarily sought by sharing physical resources, such as processors, memory, and disks in the form of virtual machines or containers among multiple users, i.e., workload consolidation. However, the reality is co-located applications in these virtual platforms compete for resources and interfere with each others' performance, resulting in performance variability/degradation. In this paper, we present the contentionaware resource allocation (CARA) solution, which optimizes data center efficiency. It is essentially devised based on a model predictive control that enables to make judicious consolidation decisions with future system states. CARA consolidates workloads explicitly taking into account the correlation between shared and isolated resource usage patterns. Based on our experimental results, CARA improves the overall resource utilization by 32%, without a significant impact on the quality-of-service (QoS) enforcement level. Such improvement results in a fewer number of active servers and in turn contributes to an overall energy saving by 33%.

UR - http://www.scopus.com/inward/record.url?scp=85010366247&partnerID=8YFLogxK

U2 - 10.1109/MASCOTS.2016.59

DO - 10.1109/MASCOTS.2016.59

M3 - Conference proceeding contribution

AN - SCOPUS:85010366247

SP - 277

EP - 282

BT - Proceedings - 2016 IEEE 24th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016

PB - Institute of Electrical and Electronics Engineers (IEEE)

CY - Piscataway, NJ

T2 - 24th IEEE International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016

Y2 - 19 September 2016 through 21 September 2016

ER -

Hoseinyfarahabady MR, Lee YC, Zomaya AY, Tari Z, Song A. A model predictive controller for contention-aware resource allocation in virtualized data centers. In Proceedings - 2016 IEEE 24th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, MASCOTS 2016. Piscataway, NJ: Institute of Electrical and Electronics Engineers (IEEE). 2016. p. 277-282. 7774590 doi: 10.1109/MASCOTS.2016.59

A model predictive controller for contention-aware resource allocation in virtualized data centers

Abstract

Other

Access to Document

Other files and links

Fingerprint

Cite this