Experimental assessment of traffic prediction assisted data center network reconfiguration method

Xiaotao Guo; Xuwei Xue; Fulong Yan; Bitao Pan; Georgios Exarchakos; Nicola Calabretta

doi:10.1109/ECOC52684.2021.9606079

Experimental assessment of traffic prediction assisted data center network reconfiguration method

Xiaotao Guo
, Xuwei Xue
, Fulong Yan
, Bitao Pan
, Georgios Exarchakos
, Nicola Calabretta

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

2 Citations (Scopus)

2 Downloads (Pure)

Abstract

We experimentally demonstrate a traffic prediction assisted network reconfiguration method (TPANR) for data center networks based on deep reinforcement learning (DRL). Traffic prediction model performs the lowest MSE of 2.64E-4. Exploiting one-step ahead traffic prediction and DRL-based automatic network reconfiguration, TPANR achieves 17.3% latency improvement.

Original language	English
Title of host publication	2021 European Conference on Optical Communication (ECOC)
Publisher	Institute of Electrical and Electronics Engineers
Number of pages	4
ISBN (Electronic)	978-1-6654-3868-1
DOIs	https://doi.org/10.1109/ECOC52684.2021.9606079
Publication status	Published - 22 Nov 2021
Event	47th European Conference on Optical Communications, ECOC 2021 - Bordeaux, France Duration: 13 Sept 2021 → 16 Sept 2021 Conference number: 47

Conference

Conference	47th European Conference on Optical Communications, ECOC 2021
Abbreviated title	ECOC 2021
Country/Territory	France
City	Bordeaux
Period	13/09/21 → 16/09/21
Other	European Conference and Exhibition on Optical Communications

Access to Document

10.1109/ECOC52684.2021.9606079

Cite this

@inproceedings{b2394543757244a38e512f84b42aa64b,

title = "Experimental assessment of traffic prediction assisted data center network reconfiguration method",

abstract = "We experimentally demonstrate a traffic prediction assisted network reconfiguration method (TPANR) for data center networks based on deep reinforcement learning (DRL). Traffic prediction model performs the lowest MSE of 2.64E-4. Exploiting one-step ahead traffic prediction and DRL-based automatic network reconfiguration, TPANR achieves 17.3\% latency improvement.",

author = "Xiaotao Guo and Xuwei Xue and Fulong Yan and Bitao Pan and Georgios Exarchakos and Nicola Calabretta",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE.; 47th European Conference on Optical Communications, ECOC 2021, ECOC 2021 ; Conference date: 13-09-2021 Through 16-09-2021",

year = "2021",

month = nov,

day = "22",

doi = "10.1109/ECOC52684.2021.9606079",

language = "English",

booktitle = "2021 European Conference on Optical Communication (ECOC)",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

}

Guo, X, Xue, X, Yan, F, Pan, B, Exarchakos, G & Calabretta, N 2021, Experimental assessment of traffic prediction assisted data center network reconfiguration method. in 2021 European Conference on Optical Communication (ECOC)., 9606079, Institute of Electrical and Electronics Engineers, 47th European Conference on Optical Communications, ECOC 2021, Bordeaux, France, 13/09/21. https://doi.org/10.1109/ECOC52684.2021.9606079

Experimental assessment of traffic prediction assisted data center network reconfiguration method. / Guo, Xiaotao; Xue, Xuwei; Yan, Fulong et al.
2021 European Conference on Optical Communication (ECOC). Institute of Electrical and Electronics Engineers, 2021. 9606079.

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

TY - GEN

T1 - Experimental assessment of traffic prediction assisted data center network reconfiguration method

AU - Guo, Xiaotao

AU - Xue, Xuwei

AU - Yan, Fulong

AU - Pan, Bitao

AU - Exarchakos, Georgios

AU - Calabretta, Nicola

N1 - Conference code: 47

PY - 2021/11/22

Y1 - 2021/11/22

N2 - We experimentally demonstrate a traffic prediction assisted network reconfiguration method (TPANR) for data center networks based on deep reinforcement learning (DRL). Traffic prediction model performs the lowest MSE of 2.64E-4. Exploiting one-step ahead traffic prediction and DRL-based automatic network reconfiguration, TPANR achieves 17.3% latency improvement.

AB - We experimentally demonstrate a traffic prediction assisted network reconfiguration method (TPANR) for data center networks based on deep reinforcement learning (DRL). Traffic prediction model performs the lowest MSE of 2.64E-4. Exploiting one-step ahead traffic prediction and DRL-based automatic network reconfiguration, TPANR achieves 17.3% latency improvement.

UR - https://www.scopus.com/pages/publications/85123220218

U2 - 10.1109/ECOC52684.2021.9606079

DO - 10.1109/ECOC52684.2021.9606079

M3 - Conference contribution

AN - SCOPUS:85123220218

BT - 2021 European Conference on Optical Communication (ECOC)

PB - Institute of Electrical and Electronics Engineers

T2 - 47th European Conference on Optical Communications, ECOC 2021

Y2 - 13 September 2021 through 16 September 2021

ER -

Experimental assessment of traffic prediction assisted data center network reconfiguration method

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this