Balanced fair resource sharing in computer clusters

Thomas Bonald; Céline Comte

doi:10.1016/j.peva.2017.08.006

Balanced fair resource sharing in computer clusters

Thomas Bonald
, Céline Comte

Research output: Contribution to journal › Article › peer-review

Abstract

We represent a computer cluster as a multi-server queue with some arbitrary graph of compatibilities between jobs and servers. Each server processes its jobs sequentially in FCFS order. The service rate of a job at any given time is the sum of the service rates of all servers processing this job. We show that the corresponding queue is quasi-reversible and use this property to design a scheduling algorithm achieving balanced fair sharing of the computing resources.

Original language	English
Pages (from-to)	70-83
Number of pages	14
Journal	Performance Evaluation
Volume	116
DOIs	https://doi.org/10.1016/j.peva.2017.08.006
Publication status	Published - 1 Nov 2017
Externally published	Yes

Keywords

Balanced fairness
Multi-server queues
Order independent queues
Parallel processing
Whittle networks

Access to Document

10.1016/j.peva.2017.08.006

Cite this

@article{dee5778d86f04ed087413c70f5a50ad8,

title = "Balanced fair resource sharing in computer clusters",

abstract = "We represent a computer cluster as a multi-server queue with some arbitrary graph of compatibilities between jobs and servers. Each server processes its jobs sequentially in FCFS order. The service rate of a job at any given time is the sum of the service rates of all servers processing this job. We show that the corresponding queue is quasi-reversible and use this property to design a scheduling algorithm achieving balanced fair sharing of the computing resources.",

keywords = "Balanced fairness, Multi-server queues, Order independent queues, Parallel processing, Whittle networks",

author = "Thomas Bonald and C{\'e}line Comte",

note = "Publisher Copyright: {\textcopyright} 2017 Elsevier B.V.",

year = "2017",

month = nov,

day = "1",

doi = "10.1016/j.peva.2017.08.006",

language = "English",

volume = "116",

pages = "70--83",

journal = "Performance Evaluation",

issn = "0166-5316",

}

TY - JOUR

T1 - Balanced fair resource sharing in computer clusters

AU - Bonald, Thomas

AU - Comte, Céline

PY - 2017/11/1

Y1 - 2017/11/1

N2 - We represent a computer cluster as a multi-server queue with some arbitrary graph of compatibilities between jobs and servers. Each server processes its jobs sequentially in FCFS order. The service rate of a job at any given time is the sum of the service rates of all servers processing this job. We show that the corresponding queue is quasi-reversible and use this property to design a scheduling algorithm achieving balanced fair sharing of the computing resources.

AB - We represent a computer cluster as a multi-server queue with some arbitrary graph of compatibilities between jobs and servers. Each server processes its jobs sequentially in FCFS order. The service rate of a job at any given time is the sum of the service rates of all servers processing this job. We show that the corresponding queue is quasi-reversible and use this property to design a scheduling algorithm achieving balanced fair sharing of the computing resources.

KW - Balanced fairness

KW - Multi-server queues

KW - Order independent queues

KW - Parallel processing

KW - Whittle networks

U2 - 10.1016/j.peva.2017.08.006

DO - 10.1016/j.peva.2017.08.006

M3 - Article

AN - SCOPUS:85029487070

SN - 0166-5316

VL - 116

SP - 70

EP - 83

JO - Performance Evaluation

JF - Performance Evaluation

ER -

Balanced fair resource sharing in computer clusters

Abstract

Keywords

Access to Document

Fingerprint

Cite this