Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows

Detalhes bibliográficos
Autor(a) principal: BEZERRA, Jeandro de Mesquita
Data de Publicação: 2020
Tipo de documento: Tese
Idioma: eng
Título da fonte: Repositório Institucional da UFPE
dARK ID: ark:/64986/0013000015w0d
Texto Completo: https://repositorio.ufpe.br/handle/123456789/40068
Resumo: Applications and services hosted in large Data Centers account for most of the increase in Internet traffic. Data Center Networks (DCNs) are often designed with a fat-tree topology, allowing multiple paths between any two servers. The most widely adopted solution for flow routing in DCNs is equal-cost multipath (ECMP), which can cause link performance degra dation due to the possible occurrence of hash collisions in the presence of flows with many gigabytes of data, called elephants. Such collisions can result in packet discard, which generates packet retransmission, causes additional latency, and further degrades link performance. This thesis proposes a hybrid prediction model by combining aspects of the FARIMA and the Recur rent Neural Network (FARIMA-RNN) models to predict elephant flows on a short-term basis. Besides, we implement an SDN solution based on a randomized rounding heuristic, named RDRH, to schedule elephant flows in DCNs. We employ a linear programming formulation that provides in polynomial time lower bounds for balancing elephant flows. A methodology based on the rank of the prediction accuracy metrics is applied to compare the hybrid model’s performance with the ARIMA, GARCH, RBF, MLP, and LSTM models. Results show that the FARIMA-RNN model presents lower error rates than the other predictors. Furthermore, we evaluate our proposed heuristic performance on an emulated network with Mininet. The ex periments show that the RDRH solution presents a performance gain compared to the ECMP and Hedera solutions in the round-trip delay and loss metrics in two of the four evaluated scenarios.
id UFPE_cf3013c3900c7a548021af41a667e016
oai_identifier_str oai:repositorio.ufpe.br:123456789/40068
network_acronym_str UFPE
network_name_str Repositório Institucional da UFPE
repository_id_str 2221
spelling BEZERRA, Jeandro de Mesquitahttp://lattes.cnpq.br/8443091736542183http://lattes.cnpq.br/9838400375894439CAMPELO, Divanilson Rodrigo de Sousa2021-05-13T17:06:47Z2021-05-13T17:06:47Z2020-11-25BEZERRA, Jeandro de Mesquita. Traffic engineering in data center networks: prediction and scheduling via randomized rounding for elephant flows. 2020. Tese (Doutorado em Ciências da Computação) - Universidade Federal de Pernambuco, Recife, 2020.https://repositorio.ufpe.br/handle/123456789/40068ark:/64986/0013000015w0dApplications and services hosted in large Data Centers account for most of the increase in Internet traffic. Data Center Networks (DCNs) are often designed with a fat-tree topology, allowing multiple paths between any two servers. The most widely adopted solution for flow routing in DCNs is equal-cost multipath (ECMP), which can cause link performance degra dation due to the possible occurrence of hash collisions in the presence of flows with many gigabytes of data, called elephants. Such collisions can result in packet discard, which generates packet retransmission, causes additional latency, and further degrades link performance. This thesis proposes a hybrid prediction model by combining aspects of the FARIMA and the Recur rent Neural Network (FARIMA-RNN) models to predict elephant flows on a short-term basis. Besides, we implement an SDN solution based on a randomized rounding heuristic, named RDRH, to schedule elephant flows in DCNs. We employ a linear programming formulation that provides in polynomial time lower bounds for balancing elephant flows. A methodology based on the rank of the prediction accuracy metrics is applied to compare the hybrid model’s performance with the ARIMA, GARCH, RBF, MLP, and LSTM models. Results show that the FARIMA-RNN model presents lower error rates than the other predictors. Furthermore, we evaluate our proposed heuristic performance on an emulated network with Mininet. The ex periments show that the RDRH solution presents a performance gain compared to the ECMP and Hedera solutions in the round-trip delay and loss metrics in two of the four evaluated scenarios.FUNCAPAplicações e serviços hospedados em grandes data centers ocasionam um aumento no volume de tráfego da Internet. Redes de data center (DCNs) são frequentemente projetadas com a topologia fat-tree que permitem múltiplos caminhos entre quaisquer dois servidores. A solução mais adotada para roteamento de fluxos em DCNs é o ECMP (equal-cost multipath), que pode causar degradação do desempenho no enlace devido à possibilidade de ocorrência de colisões de hash na presença de fluxos com muitos gigabytes, chamados de elefante. Tais colisões podem ocasionar descarte de pacotes que geram retransmissões causando atrasos adicionais e degradam o desempenho do enlace. Esta tese propõe um modelo híbrido de predição combinando aspectos dos modelos FARIMA e de Redes Neurais Recorrentes, chamado de (FARIMA-RNN), para prever fluxos elefante em curto período. Além disso, implementamos uma solução SDN baseada em uma heurística de arredondamento probabilístico, denominada de RDRH, para escalonar fluxos elefante em DCNs. Uma metodologia baseada no rank da métrica de acurácia de previsão é aplicada para comparar o desempenho do modelo híbrido com os modelos ARIMA, GARCH, RBF, MLP e LSTM. Os resultados mostram que o modelo FARIMA-RNN apresenta taxas de erro menores que os demais modelos. A heurística proposta foi avaliada em uma DCN emulada com o Mininet. Os experimentos mostram que a solução RDRH apresenta ganho de desempenho comparado com as soluções ECMP e Hedera nas métricas atraso de ida e volta e perda em dois dos quatro cenários avaliadosengUniversidade Federal de PernambucoPrograma de Pos Graduacao em Ciencia da ComputacaoUFPEBrasilAttribution-NonCommercial-NoDerivs 3.0 Brazilhttp://creativecommons.org/licenses/by-nc-nd/3.0/br/info:eu-repo/semantics/openAccessRedes de computadoresAvaliação de desempenhoTraffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flowsinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/doctoralThesisdoutoradoreponame:Repositório Institucional da UFPEinstname:Universidade Federal de Pernambuco (UFPE)instacron:UFPETEXTTESE Jeandro de Mesquita Bezerra.pdf.txtTESE Jeandro de Mesquita Bezerra.pdf.txtExtracted texttext/plain194787https://repositorio.ufpe.br/bitstream/123456789/40068/4/TESE%20Jeandro%20de%20Mesquita%20Bezerra.pdf.txt6f3a69ede3adebc1da5f74ff87473eb1MD54THUMBNAILTESE Jeandro de Mesquita Bezerra.pdf.jpgTESE Jeandro de Mesquita Bezerra.pdf.jpgGenerated Thumbnailimage/jpeg1214https://repositorio.ufpe.br/bitstream/123456789/40068/5/TESE%20Jeandro%20de%20Mesquita%20Bezerra.pdf.jpgd75dc33e8396cd82be6adeabdbaa6eb8MD55CC-LICENSElicense_rdflicense_rdfapplication/rdf+xml; charset=utf-8811https://repositorio.ufpe.br/bitstream/123456789/40068/2/license_rdfe39d27027a6cc9cb039ad269a5db8e34MD52ORIGINALTESE Jeandro de Mesquita Bezerra.pdfTESE Jeandro de Mesquita Bezerra.pdfapplication/pdf3152649https://repositorio.ufpe.br/bitstream/123456789/40068/1/TESE%20Jeandro%20de%20Mesquita%20Bezerra.pdf19befb9a25bca6cba7af6b3fafcf078eMD51LICENSElicense.txtlicense.txttext/plain; charset=utf-82310https://repositorio.ufpe.br/bitstream/123456789/40068/3/license.txtbd573a5ca8288eb7272482765f819534MD53123456789/400682021-05-14 02:17:03.442oai:repositorio.ufpe.br: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ório InstitucionalPUBhttps://repositorio.ufpe.br/oai/requestattena@ufpe.bropendoar:22212021-05-14T05:17:03Repositório Institucional da UFPE - Universidade Federal de Pernambuco (UFPE)false
dc.title.pt_BR.fl_str_mv Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows
title Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows
spellingShingle Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows
BEZERRA, Jeandro de Mesquita
Redes de computadores
Avaliação de desempenho
title_short Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows
title_full Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows
title_fullStr Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows
title_full_unstemmed Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows
title_sort Traffic engineering in data center networks : prediction and scheduling via randomized rounding for elephant flows
author BEZERRA, Jeandro de Mesquita
author_facet BEZERRA, Jeandro de Mesquita
author_role author
dc.contributor.authorLattes.pt_BR.fl_str_mv http://lattes.cnpq.br/8443091736542183
dc.contributor.advisorLattes.pt_BR.fl_str_mv http://lattes.cnpq.br/9838400375894439
dc.contributor.author.fl_str_mv BEZERRA, Jeandro de Mesquita
dc.contributor.advisor1.fl_str_mv CAMPELO, Divanilson Rodrigo de Sousa
contributor_str_mv CAMPELO, Divanilson Rodrigo de Sousa
dc.subject.por.fl_str_mv Redes de computadores
Avaliação de desempenho
topic Redes de computadores
Avaliação de desempenho
description Applications and services hosted in large Data Centers account for most of the increase in Internet traffic. Data Center Networks (DCNs) are often designed with a fat-tree topology, allowing multiple paths between any two servers. The most widely adopted solution for flow routing in DCNs is equal-cost multipath (ECMP), which can cause link performance degra dation due to the possible occurrence of hash collisions in the presence of flows with many gigabytes of data, called elephants. Such collisions can result in packet discard, which generates packet retransmission, causes additional latency, and further degrades link performance. This thesis proposes a hybrid prediction model by combining aspects of the FARIMA and the Recur rent Neural Network (FARIMA-RNN) models to predict elephant flows on a short-term basis. Besides, we implement an SDN solution based on a randomized rounding heuristic, named RDRH, to schedule elephant flows in DCNs. We employ a linear programming formulation that provides in polynomial time lower bounds for balancing elephant flows. A methodology based on the rank of the prediction accuracy metrics is applied to compare the hybrid model’s performance with the ARIMA, GARCH, RBF, MLP, and LSTM models. Results show that the FARIMA-RNN model presents lower error rates than the other predictors. Furthermore, we evaluate our proposed heuristic performance on an emulated network with Mininet. The ex periments show that the RDRH solution presents a performance gain compared to the ECMP and Hedera solutions in the round-trip delay and loss metrics in two of the four evaluated scenarios.
publishDate 2020
dc.date.issued.fl_str_mv 2020-11-25
dc.date.accessioned.fl_str_mv 2021-05-13T17:06:47Z
dc.date.available.fl_str_mv 2021-05-13T17:06:47Z
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/doctoralThesis
format doctoralThesis
status_str publishedVersion
dc.identifier.citation.fl_str_mv BEZERRA, Jeandro de Mesquita. Traffic engineering in data center networks: prediction and scheduling via randomized rounding for elephant flows. 2020. Tese (Doutorado em Ciências da Computação) - Universidade Federal de Pernambuco, Recife, 2020.
dc.identifier.uri.fl_str_mv https://repositorio.ufpe.br/handle/123456789/40068
dc.identifier.dark.fl_str_mv ark:/64986/0013000015w0d
identifier_str_mv BEZERRA, Jeandro de Mesquita. Traffic engineering in data center networks: prediction and scheduling via randomized rounding for elephant flows. 2020. Tese (Doutorado em Ciências da Computação) - Universidade Federal de Pernambuco, Recife, 2020.
ark:/64986/0013000015w0d
url https://repositorio.ufpe.br/handle/123456789/40068
dc.language.iso.fl_str_mv eng
language eng
dc.rights.driver.fl_str_mv Attribution-NonCommercial-NoDerivs 3.0 Brazil
http://creativecommons.org/licenses/by-nc-nd/3.0/br/
info:eu-repo/semantics/openAccess
rights_invalid_str_mv Attribution-NonCommercial-NoDerivs 3.0 Brazil
http://creativecommons.org/licenses/by-nc-nd/3.0/br/
eu_rights_str_mv openAccess
dc.publisher.none.fl_str_mv Universidade Federal de Pernambuco
dc.publisher.program.fl_str_mv Programa de Pos Graduacao em Ciencia da Computacao
dc.publisher.initials.fl_str_mv UFPE
dc.publisher.country.fl_str_mv Brasil
publisher.none.fl_str_mv Universidade Federal de Pernambuco
dc.source.none.fl_str_mv reponame:Repositório Institucional da UFPE
instname:Universidade Federal de Pernambuco (UFPE)
instacron:UFPE
instname_str Universidade Federal de Pernambuco (UFPE)
instacron_str UFPE
institution UFPE
reponame_str Repositório Institucional da UFPE
collection Repositório Institucional da UFPE
bitstream.url.fl_str_mv https://repositorio.ufpe.br/bitstream/123456789/40068/4/TESE%20Jeandro%20de%20Mesquita%20Bezerra.pdf.txt
https://repositorio.ufpe.br/bitstream/123456789/40068/5/TESE%20Jeandro%20de%20Mesquita%20Bezerra.pdf.jpg
https://repositorio.ufpe.br/bitstream/123456789/40068/2/license_rdf
https://repositorio.ufpe.br/bitstream/123456789/40068/1/TESE%20Jeandro%20de%20Mesquita%20Bezerra.pdf
https://repositorio.ufpe.br/bitstream/123456789/40068/3/license.txt
bitstream.checksum.fl_str_mv 6f3a69ede3adebc1da5f74ff87473eb1
d75dc33e8396cd82be6adeabdbaa6eb8
e39d27027a6cc9cb039ad269a5db8e34
19befb9a25bca6cba7af6b3fafcf078e
bd573a5ca8288eb7272482765f819534
bitstream.checksumAlgorithm.fl_str_mv MD5
MD5
MD5
MD5
MD5
repository.name.fl_str_mv Repositório Institucional da UFPE - Universidade Federal de Pernambuco (UFPE)
repository.mail.fl_str_mv attena@ufpe.br
_version_ 1815173013757231104