Analytical model and performance analysis of a network interface card

N. Cherukuri, G. B. Kandiraju, N. Gautam, Anand Sivasubramaniam

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

One important issue for practitioners and academicians is that there are almost no platforms based on analytical models for testing the impact of various architectural and design modifications for intelligent network interface cards (NICs). Simulations are typically time consuming, especially for experimenting with different scenarios and what-if analysis. In this research, we study the performance of an NIC called Myrinet developed by Myricom. We develop an open queueing network model to predict its performance and compare the analytical results with the simulations. The reason that there are very few analytical models is that the performance-analysis problem is enormously complex. In particular, the problem is a combination of (a) multiclass queueing network with class switching, (b) polling system with limited service discipline, and (c) finite-capacity queues with blocking. These three issues have been treated only in isolation in the literature. However, the problem becomes much harder when all three issues are simultaneously present. One of the key contributions of this work is an analytical approximation of this complex system. From an analytical modelling standpoint, we observe that making simplifying assumptions to analyze nodes that are not bottlenecks does not greatly affect performance. The main findings of this research are the bottlenecks of the queueing network, utilizations of the various nodes, and performance measures such as the expected delay. The model as well as findings can be used to test the performance impact of various enhancements to the operation of NICs.

Original languageEnglish (US)
Pages (from-to)179-189
Number of pages11
JournalInternational Journal of Modelling and Simulation
Volume24
Issue number3
DOIs
StatePublished - Jan 1 2004

Fingerprint

Queueing networks
Model Analysis
Analytical Model
Interfaces (computer)
Performance Analysis
Analytical models
Intelligent networks
Multiclass Queueing Networks
Polling Systems
Queueing Network Models
Finite Capacity
Analytical Modeling
Analytical Approximation
Queueing Networks
Large scale systems
Vertex of a graph
Performance Measures
Isolation
Queue
Complex Systems

All Science Journal Classification (ASJC) codes

  • Software
  • Modeling and Simulation
  • Mechanics of Materials
  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this

@article{ec0ad510371e4b0e990824c1d1de0d8d,
title = "Analytical model and performance analysis of a network interface card",
abstract = "One important issue for practitioners and academicians is that there are almost no platforms based on analytical models for testing the impact of various architectural and design modifications for intelligent network interface cards (NICs). Simulations are typically time consuming, especially for experimenting with different scenarios and what-if analysis. In this research, we study the performance of an NIC called Myrinet developed by Myricom. We develop an open queueing network model to predict its performance and compare the analytical results with the simulations. The reason that there are very few analytical models is that the performance-analysis problem is enormously complex. In particular, the problem is a combination of (a) multiclass queueing network with class switching, (b) polling system with limited service discipline, and (c) finite-capacity queues with blocking. These three issues have been treated only in isolation in the literature. However, the problem becomes much harder when all three issues are simultaneously present. One of the key contributions of this work is an analytical approximation of this complex system. From an analytical modelling standpoint, we observe that making simplifying assumptions to analyze nodes that are not bottlenecks does not greatly affect performance. The main findings of this research are the bottlenecks of the queueing network, utilizations of the various nodes, and performance measures such as the expected delay. The model as well as findings can be used to test the performance impact of various enhancements to the operation of NICs.",
author = "N. Cherukuri and Kandiraju, {G. B.} and N. Gautam and Anand Sivasubramaniam",
year = "2004",
month = "1",
day = "1",
doi = "10.1080/02286203.2004.11442302",
language = "English (US)",
volume = "24",
pages = "179--189",
journal = "International Journal of Modelling and Simulation",
issn = "0228-6203",
publisher = "ACTA Press",
number = "3",

}

Analytical model and performance analysis of a network interface card. / Cherukuri, N.; Kandiraju, G. B.; Gautam, N.; Sivasubramaniam, Anand.

In: International Journal of Modelling and Simulation, Vol. 24, No. 3, 01.01.2004, p. 179-189.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Analytical model and performance analysis of a network interface card

AU - Cherukuri, N.

AU - Kandiraju, G. B.

AU - Gautam, N.

AU - Sivasubramaniam, Anand

PY - 2004/1/1

Y1 - 2004/1/1

N2 - One important issue for practitioners and academicians is that there are almost no platforms based on analytical models for testing the impact of various architectural and design modifications for intelligent network interface cards (NICs). Simulations are typically time consuming, especially for experimenting with different scenarios and what-if analysis. In this research, we study the performance of an NIC called Myrinet developed by Myricom. We develop an open queueing network model to predict its performance and compare the analytical results with the simulations. The reason that there are very few analytical models is that the performance-analysis problem is enormously complex. In particular, the problem is a combination of (a) multiclass queueing network with class switching, (b) polling system with limited service discipline, and (c) finite-capacity queues with blocking. These three issues have been treated only in isolation in the literature. However, the problem becomes much harder when all three issues are simultaneously present. One of the key contributions of this work is an analytical approximation of this complex system. From an analytical modelling standpoint, we observe that making simplifying assumptions to analyze nodes that are not bottlenecks does not greatly affect performance. The main findings of this research are the bottlenecks of the queueing network, utilizations of the various nodes, and performance measures such as the expected delay. The model as well as findings can be used to test the performance impact of various enhancements to the operation of NICs.

AB - One important issue for practitioners and academicians is that there are almost no platforms based on analytical models for testing the impact of various architectural and design modifications for intelligent network interface cards (NICs). Simulations are typically time consuming, especially for experimenting with different scenarios and what-if analysis. In this research, we study the performance of an NIC called Myrinet developed by Myricom. We develop an open queueing network model to predict its performance and compare the analytical results with the simulations. The reason that there are very few analytical models is that the performance-analysis problem is enormously complex. In particular, the problem is a combination of (a) multiclass queueing network with class switching, (b) polling system with limited service discipline, and (c) finite-capacity queues with blocking. These three issues have been treated only in isolation in the literature. However, the problem becomes much harder when all three issues are simultaneously present. One of the key contributions of this work is an analytical approximation of this complex system. From an analytical modelling standpoint, we observe that making simplifying assumptions to analyze nodes that are not bottlenecks does not greatly affect performance. The main findings of this research are the bottlenecks of the queueing network, utilizations of the various nodes, and performance measures such as the expected delay. The model as well as findings can be used to test the performance impact of various enhancements to the operation of NICs.

UR - http://www.scopus.com/inward/record.url?scp=2942572826&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=2942572826&partnerID=8YFLogxK

U2 - 10.1080/02286203.2004.11442302

DO - 10.1080/02286203.2004.11442302

M3 - Article

VL - 24

SP - 179

EP - 189

JO - International Journal of Modelling and Simulation

JF - International Journal of Modelling and Simulation

SN - 0228-6203

IS - 3

ER -