Combining hardware and simulation for datacenter scaling studies

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedings – Annual report year: 2017Researchpeer-review

Standard

Combining hardware and simulation for datacenter scaling studies. / Ruepp, Sarah Renée; Pilimon, Artur; Thrane, Jakob; Galili, Michael; Berger, Michael Stübert; Dittmann, Lars.

2017 International Conference on Optical Network Design and Modeling. IEEE, 2017. p. 1-6.

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedings – Annual report year: 2017Researchpeer-review

Harvard

Ruepp, SR, Pilimon, A, Thrane, J, Galili, M, Berger, MS & Dittmann, L 2017, Combining hardware and simulation for datacenter scaling studies. in 2017 International Conference on Optical Network Design and Modeling. IEEE, pp. 1-6, 2017 International Conference on Optical Network Design and Modeling, Budapest, Hungary, 15/05/2017. https://doi.org/10.23919/ONDM.2017.7958554

APA

CBE

MLA

Ruepp, Sarah Renée et al. "Combining hardware and simulation for datacenter scaling studies". 2017 International Conference on Optical Network Design and Modeling. IEEE. 2017, 1-6. https://doi.org/10.23919/ONDM.2017.7958554

Vancouver

Author

Bibtex

@inproceedings{4d63342cdf014f2e967000d83134f845,
title = "Combining hardware and simulation for datacenter scaling studies",
abstract = "Datacenter networks are becoming crucial foundations for our information technology based society. However, commercial datacenter infrastructure is often unavailable to researchers for conducting experiments. In this work, we therefore elaborate on the possibility of combining commercial hardware and simulation to illustrate the scalability and performance of datacenter networks. We simulate a Datacenter network and interconnect it with real world traffic generation hardware. Analysis of the introduced packet conversion and virtual queueing delays shows that the conversion efficiency is at the order of a few microseconds, but the virtual queuing may have significant implication on the performance analysis results.",
keywords = "Delays, Analytical models, Computational modeling, Hardware, Servers, Computer architecture, Workstations, Datacenter, Simulation, System-in-the-loop",
author = "Ruepp, {Sarah Ren{\'e}e} and Artur Pilimon and Jakob Thrane and Michael Galili and Berger, {Michael St{\"u}bert} and Lars Dittmann",
year = "2017",
doi = "10.23919/ONDM.2017.7958554",
language = "English",
isbn = "9783901882937",
pages = "1--6",
booktitle = "2017 International Conference on Optical Network Design and Modeling",
publisher = "IEEE",
address = "United States",

}

RIS

TY - GEN

T1 - Combining hardware and simulation for datacenter scaling studies

AU - Ruepp, Sarah Renée

AU - Pilimon, Artur

AU - Thrane, Jakob

AU - Galili, Michael

AU - Berger, Michael Stübert

AU - Dittmann, Lars

PY - 2017

Y1 - 2017

N2 - Datacenter networks are becoming crucial foundations for our information technology based society. However, commercial datacenter infrastructure is often unavailable to researchers for conducting experiments. In this work, we therefore elaborate on the possibility of combining commercial hardware and simulation to illustrate the scalability and performance of datacenter networks. We simulate a Datacenter network and interconnect it with real world traffic generation hardware. Analysis of the introduced packet conversion and virtual queueing delays shows that the conversion efficiency is at the order of a few microseconds, but the virtual queuing may have significant implication on the performance analysis results.

AB - Datacenter networks are becoming crucial foundations for our information technology based society. However, commercial datacenter infrastructure is often unavailable to researchers for conducting experiments. In this work, we therefore elaborate on the possibility of combining commercial hardware and simulation to illustrate the scalability and performance of datacenter networks. We simulate a Datacenter network and interconnect it with real world traffic generation hardware. Analysis of the introduced packet conversion and virtual queueing delays shows that the conversion efficiency is at the order of a few microseconds, but the virtual queuing may have significant implication on the performance analysis results.

KW - Delays

KW - Analytical models

KW - Computational modeling

KW - Hardware

KW - Servers

KW - Computer architecture

KW - Workstations

KW - Datacenter

KW - Simulation

KW - System-in-the-loop

U2 - 10.23919/ONDM.2017.7958554

DO - 10.23919/ONDM.2017.7958554

M3 - Article in proceedings

SN - 9783901882937

SP - 1

EP - 6

BT - 2017 International Conference on Optical Network Design and Modeling

PB - IEEE

ER -