Analytical modelling of networks in multicomputer systems under bursty and batch arrival traffic

Yulei Wu, Geyong Min, Mohamed Ould-Khaoua, Hao Yin, Lan Wang

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

The hypercube and torus are two important message-passing network architectures of high-performance multicomputers. Analytical models of multicomputer networks under the non-bursty Poisson traffic have been widely reported. Motivated by the convincing evidence of bursty and batch arrival nature of traffic generated by many real-world parallel applications in high-performance computing environments, we develop a new and concise analytical model in this paper for hypercube and torus networks in the presence of batch message arrivals modelled by the compound Poisson process with geometrically distributed batch sizes. The average degree of virtual channel multiplexing is derived by employing a Markov chain which can capture the batch arrival nature. An attractive advantage of the model is its constant computation complexity independent of the network size. The accuracy of the analytical performance results is validated against those obtained from simulation experiments of an actual system.

Original languageEnglish
Pages (from-to)115-130
Number of pages16
JournalJournal of Supercomputing
Volume51
Issue number2
DOIs
Publication statusPublished - Feb 2010

Fingerprint

Batch Arrivals
Analytical Modeling
Multicomputers
Analytical models
Traffic
Hypercube
Analytical Model
Batch
Torus
Message passing
High Performance
Network architecture
Multiplexing
Markov processes
Virtual Channel
Compound Poisson Process
Parallel Applications
Network Architecture
Message Passing
Real-world Applications

Keywords

  • Adaptive routing
  • Compound Poisson process
  • Generalised exponential distribution
  • Interconnection networks
  • Performance modelling

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software
  • Information Systems
  • Theoretical Computer Science

Cite this

Analytical modelling of networks in multicomputer systems under bursty and batch arrival traffic. / Wu, Yulei; Min, Geyong; Ould-Khaoua, Mohamed; Yin, Hao; Wang, Lan.

In: Journal of Supercomputing, Vol. 51, No. 2, 02.2010, p. 115-130.

Research output: Contribution to journalArticle

@article{c69a2dc874f94eaa8fe02d10f04e79e9,
title = "Analytical modelling of networks in multicomputer systems under bursty and batch arrival traffic",
abstract = "The hypercube and torus are two important message-passing network architectures of high-performance multicomputers. Analytical models of multicomputer networks under the non-bursty Poisson traffic have been widely reported. Motivated by the convincing evidence of bursty and batch arrival nature of traffic generated by many real-world parallel applications in high-performance computing environments, we develop a new and concise analytical model in this paper for hypercube and torus networks in the presence of batch message arrivals modelled by the compound Poisson process with geometrically distributed batch sizes. The average degree of virtual channel multiplexing is derived by employing a Markov chain which can capture the batch arrival nature. An attractive advantage of the model is its constant computation complexity independent of the network size. The accuracy of the analytical performance results is validated against those obtained from simulation experiments of an actual system.",
keywords = "Adaptive routing, Compound Poisson process, Generalised exponential distribution, Interconnection networks, Performance modelling",
author = "Yulei Wu and Geyong Min and Mohamed Ould-Khaoua and Hao Yin and Lan Wang",
year = "2010",
month = "2",
doi = "10.1007/s11227-009-0265-x",
language = "English",
volume = "51",
pages = "115--130",
journal = "Journal of Supercomputing",
issn = "0920-8542",
publisher = "Springer Netherlands",
number = "2",

}

TY - JOUR

T1 - Analytical modelling of networks in multicomputer systems under bursty and batch arrival traffic

AU - Wu, Yulei

AU - Min, Geyong

AU - Ould-Khaoua, Mohamed

AU - Yin, Hao

AU - Wang, Lan

PY - 2010/2

Y1 - 2010/2

N2 - The hypercube and torus are two important message-passing network architectures of high-performance multicomputers. Analytical models of multicomputer networks under the non-bursty Poisson traffic have been widely reported. Motivated by the convincing evidence of bursty and batch arrival nature of traffic generated by many real-world parallel applications in high-performance computing environments, we develop a new and concise analytical model in this paper for hypercube and torus networks in the presence of batch message arrivals modelled by the compound Poisson process with geometrically distributed batch sizes. The average degree of virtual channel multiplexing is derived by employing a Markov chain which can capture the batch arrival nature. An attractive advantage of the model is its constant computation complexity independent of the network size. The accuracy of the analytical performance results is validated against those obtained from simulation experiments of an actual system.

AB - The hypercube and torus are two important message-passing network architectures of high-performance multicomputers. Analytical models of multicomputer networks under the non-bursty Poisson traffic have been widely reported. Motivated by the convincing evidence of bursty and batch arrival nature of traffic generated by many real-world parallel applications in high-performance computing environments, we develop a new and concise analytical model in this paper for hypercube and torus networks in the presence of batch message arrivals modelled by the compound Poisson process with geometrically distributed batch sizes. The average degree of virtual channel multiplexing is derived by employing a Markov chain which can capture the batch arrival nature. An attractive advantage of the model is its constant computation complexity independent of the network size. The accuracy of the analytical performance results is validated against those obtained from simulation experiments of an actual system.

KW - Adaptive routing

KW - Compound Poisson process

KW - Generalised exponential distribution

KW - Interconnection networks

KW - Performance modelling

UR - http://www.scopus.com/inward/record.url?scp=76549092217&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=76549092217&partnerID=8YFLogxK

U2 - 10.1007/s11227-009-0265-x

DO - 10.1007/s11227-009-0265-x

M3 - Article

VL - 51

SP - 115

EP - 130

JO - Journal of Supercomputing

JF - Journal of Supercomputing

SN - 0920-8542

IS - 2

ER -