Loading…

BurstBalancer: Do Less, Better Balance for Large-Scale Data Center Traffic

Layer-3 load balancing is a key topic in the networking field. It is well acknowledged that flowlet is the most promising solution because of its good trade-off between load balance and packet reordering. However, we find its one significant limitation: it makes the forwarding paths of flows unpredi...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on parallel and distributed systems 2024-06, Vol.35 (6), p.932-949
Main Authors: Liu, Zirui, Zhao, Yikai, Fan, Zhuochen, Yang, Tong, Li, Xiaodong, Zhang, Ruwen, Yang, Kaicheng, Jiang, Zihan, Zhong, Zheng, Huang, Yi, Liu, Cong, Hu, Jing, Xie, Gaogang, Cui, Bin
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites cdi_FETCH-LOGICAL-c246t-289374db34b6fa98bb1213b13ea8ebd093a3205ac561a3b2d777728d289e9063
container_end_page 949
container_issue 6
container_start_page 932
container_title IEEE transactions on parallel and distributed systems
container_volume 35
creator Liu, Zirui
Zhao, Yikai
Fan, Zhuochen
Yang, Tong
Li, Xiaodong
Zhang, Ruwen
Yang, Kaicheng
Jiang, Zihan
Zhong, Zheng
Huang, Yi
Liu, Cong
Hu, Jing
Xie, Gaogang
Cui, Bin
description Layer-3 load balancing is a key topic in the networking field. It is well acknowledged that flowlet is the most promising solution because of its good trade-off between load balance and packet reordering. However, we find its one significant limitation: it makes the forwarding paths of flows unpredictable. To address this limitation, this article presents BurstBalancer, a simple yet efficient load balancing system with a sketch, named BalanceSketch. Our design philosophy is doing less changes to keep the forwarding path of most flows fixed, which guides the design of BalanceSketch and our balance operations. We have fully implemented BurstBalancer in a small-scale testbed built with Tofino switches, and conducted both large-scale event-level (NS-2) and ESL (electronic system level) simulations. Our results show that BurstBalancer achieves 5%\sim ∼ 35% smaller FCT than LetFlow in symmetric topology and up to 30× smaller FCT in asymmetric topology, while 58× fewer flows suffer from path changing. All related codes are open-sourced at GitHub.
doi_str_mv 10.1109/TPDS.2023.3295454
format article
fullrecord <record><control><sourceid>proquest_ieee_</sourceid><recordid>TN_cdi_ieee_primary_10184046</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10184046</ieee_id><sourcerecordid>3037644018</sourcerecordid><originalsourceid>FETCH-LOGICAL-c246t-289374db34b6fa98bb1213b13ea8ebd093a3205ac561a3b2d777728d289e9063</originalsourceid><addsrcrecordid>eNpNkE9PwzAMxSMEEmPwAZA4ROJKRxwnXcKNbfxVJZDWe5S0Lto01pF0B749rbYDvtiS3_OTf4xdg5gACHtffi6WEykkTlBarbQ6YSPQ2mQSDJ72s1A6sxLsObtIaS0EKC3UiL3P9jF1M7_x24riA1-0vKCU7viMuo4iP25400Ze-PhF2bLyG-IL33k-p-2gKaNvmlV1yc4av0l0dexjVj4_lfPXrPh4eZs_FlklVd5l0licqjqgCnnjrQkBJGAAJG8o1MKiRym0r3QOHoOsp31JU_c-siLHMbs9nN3F9mdPqXPrdh-3faJDgdNcKdG_PGZwUFWxTSlS43Zx9e3jrwPhBmJuIOYGYu5IrPfcHDwrIvqnB6OEyvEPLrdk2A</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3037644018</pqid></control><display><type>article</type><title>BurstBalancer: Do Less, Better Balance for Large-Scale Data Center Traffic</title><source>IEEE Xplore (Online service)</source><creator>Liu, Zirui ; Zhao, Yikai ; Fan, Zhuochen ; Yang, Tong ; Li, Xiaodong ; Zhang, Ruwen ; Yang, Kaicheng ; Jiang, Zihan ; Zhong, Zheng ; Huang, Yi ; Liu, Cong ; Hu, Jing ; Xie, Gaogang ; Cui, Bin</creator><creatorcontrib>Liu, Zirui ; Zhao, Yikai ; Fan, Zhuochen ; Yang, Tong ; Li, Xiaodong ; Zhang, Ruwen ; Yang, Kaicheng ; Jiang, Zihan ; Zhong, Zheng ; Huang, Yi ; Liu, Cong ; Hu, Jing ; Xie, Gaogang ; Cui, Bin</creatorcontrib><description><![CDATA[Layer-3 load balancing is a key topic in the networking field. It is well acknowledged that flowlet is the most promising solution because of its good trade-off between load balance and packet reordering. However, we find its one significant limitation: it makes the forwarding paths of flows unpredictable. To address this limitation, this article presents BurstBalancer, a simple yet efficient load balancing system with a sketch, named BalanceSketch. Our design philosophy is doing less changes to keep the forwarding path of most flows fixed, which guides the design of BalanceSketch and our balance operations. We have fully implemented BurstBalancer in a small-scale testbed built with Tofino switches, and conducted both large-scale event-level (NS-2) and ESL (electronic system level) simulations. Our results show that BurstBalancer achieves 5%<inline-formula><tex-math notation="LaTeX">\sim</tex-math> <mml:math><mml:mo>∼</mml:mo></mml:math><inline-graphic xlink:href="liu-ieq1-3295454.gif"/> </inline-formula>35% smaller FCT than LetFlow in symmetric topology and up to 30× smaller FCT in asymmetric topology, while 58× fewer flows suffer from path changing. All related codes are open-sourced at GitHub.]]></description><identifier>ISSN: 1045-9219</identifier><identifier>EISSN: 1558-2183</identifier><identifier>DOI: 10.1109/TPDS.2023.3295454</identifier><identifier>CODEN: ITDSEO</identifier><language>eng</language><publisher>New York: IEEE</publisher><subject>Data center networks ; Data centers ; Electronic systems ; flowlet ; L3 load balancing ; Load balancing ; Load management ; Memory management ; Network topology ; Protocols ; sketch ; System-on-chip ; Topology</subject><ispartof>IEEE transactions on parallel and distributed systems, 2024-06, Vol.35 (6), p.932-949</ispartof><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c246t-289374db34b6fa98bb1213b13ea8ebd093a3205ac561a3b2d777728d289e9063</cites><orcidid>0000-0003-2402-5854 ; 0000-0002-3534-8045 ; 0000-0003-2495-7774 ; 0000-0003-1681-4677 ; 0000-0002-6102-9195 ; 0000-0001-6381-4026 ; 0000-0003-0042-1828 ; 0000-0002-6741-6751 ; 0000-0003-4964-1135 ; 0000-0003-1160-1058 ; 0000-0001-9062-6565</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/10184046$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,27924,27925,54796</link.rule.ids></links><search><creatorcontrib>Liu, Zirui</creatorcontrib><creatorcontrib>Zhao, Yikai</creatorcontrib><creatorcontrib>Fan, Zhuochen</creatorcontrib><creatorcontrib>Yang, Tong</creatorcontrib><creatorcontrib>Li, Xiaodong</creatorcontrib><creatorcontrib>Zhang, Ruwen</creatorcontrib><creatorcontrib>Yang, Kaicheng</creatorcontrib><creatorcontrib>Jiang, Zihan</creatorcontrib><creatorcontrib>Zhong, Zheng</creatorcontrib><creatorcontrib>Huang, Yi</creatorcontrib><creatorcontrib>Liu, Cong</creatorcontrib><creatorcontrib>Hu, Jing</creatorcontrib><creatorcontrib>Xie, Gaogang</creatorcontrib><creatorcontrib>Cui, Bin</creatorcontrib><title>BurstBalancer: Do Less, Better Balance for Large-Scale Data Center Traffic</title><title>IEEE transactions on parallel and distributed systems</title><addtitle>TPDS</addtitle><description><![CDATA[Layer-3 load balancing is a key topic in the networking field. It is well acknowledged that flowlet is the most promising solution because of its good trade-off between load balance and packet reordering. However, we find its one significant limitation: it makes the forwarding paths of flows unpredictable. To address this limitation, this article presents BurstBalancer, a simple yet efficient load balancing system with a sketch, named BalanceSketch. Our design philosophy is doing less changes to keep the forwarding path of most flows fixed, which guides the design of BalanceSketch and our balance operations. We have fully implemented BurstBalancer in a small-scale testbed built with Tofino switches, and conducted both large-scale event-level (NS-2) and ESL (electronic system level) simulations. Our results show that BurstBalancer achieves 5%<inline-formula><tex-math notation="LaTeX">\sim</tex-math> <mml:math><mml:mo>∼</mml:mo></mml:math><inline-graphic xlink:href="liu-ieq1-3295454.gif"/> </inline-formula>35% smaller FCT than LetFlow in symmetric topology and up to 30× smaller FCT in asymmetric topology, while 58× fewer flows suffer from path changing. All related codes are open-sourced at GitHub.]]></description><subject>Data center networks</subject><subject>Data centers</subject><subject>Electronic systems</subject><subject>flowlet</subject><subject>L3 load balancing</subject><subject>Load balancing</subject><subject>Load management</subject><subject>Memory management</subject><subject>Network topology</subject><subject>Protocols</subject><subject>sketch</subject><subject>System-on-chip</subject><subject>Topology</subject><issn>1045-9219</issn><issn>1558-2183</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNpNkE9PwzAMxSMEEmPwAZA4ROJKRxwnXcKNbfxVJZDWe5S0Lto01pF0B749rbYDvtiS3_OTf4xdg5gACHtffi6WEykkTlBarbQ6YSPQ2mQSDJ72s1A6sxLsObtIaS0EKC3UiL3P9jF1M7_x24riA1-0vKCU7viMuo4iP25400Ze-PhF2bLyG-IL33k-p-2gKaNvmlV1yc4av0l0dexjVj4_lfPXrPh4eZs_FlklVd5l0licqjqgCnnjrQkBJGAAJG8o1MKiRym0r3QOHoOsp31JU_c-siLHMbs9nN3F9mdPqXPrdh-3faJDgdNcKdG_PGZwUFWxTSlS43Zx9e3jrwPhBmJuIOYGYu5IrPfcHDwrIvqnB6OEyvEPLrdk2A</recordid><startdate>20240601</startdate><enddate>20240601</enddate><creator>Liu, Zirui</creator><creator>Zhao, Yikai</creator><creator>Fan, Zhuochen</creator><creator>Yang, Tong</creator><creator>Li, Xiaodong</creator><creator>Zhang, Ruwen</creator><creator>Yang, Kaicheng</creator><creator>Jiang, Zihan</creator><creator>Zhong, Zheng</creator><creator>Huang, Yi</creator><creator>Liu, Cong</creator><creator>Hu, Jing</creator><creator>Xie, Gaogang</creator><creator>Cui, Bin</creator><general>IEEE</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>7SP</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><orcidid>https://orcid.org/0000-0003-2402-5854</orcidid><orcidid>https://orcid.org/0000-0002-3534-8045</orcidid><orcidid>https://orcid.org/0000-0003-2495-7774</orcidid><orcidid>https://orcid.org/0000-0003-1681-4677</orcidid><orcidid>https://orcid.org/0000-0002-6102-9195</orcidid><orcidid>https://orcid.org/0000-0001-6381-4026</orcidid><orcidid>https://orcid.org/0000-0003-0042-1828</orcidid><orcidid>https://orcid.org/0000-0002-6741-6751</orcidid><orcidid>https://orcid.org/0000-0003-4964-1135</orcidid><orcidid>https://orcid.org/0000-0003-1160-1058</orcidid><orcidid>https://orcid.org/0000-0001-9062-6565</orcidid></search><sort><creationdate>20240601</creationdate><title>BurstBalancer: Do Less, Better Balance for Large-Scale Data Center Traffic</title><author>Liu, Zirui ; Zhao, Yikai ; Fan, Zhuochen ; Yang, Tong ; Li, Xiaodong ; Zhang, Ruwen ; Yang, Kaicheng ; Jiang, Zihan ; Zhong, Zheng ; Huang, Yi ; Liu, Cong ; Hu, Jing ; Xie, Gaogang ; Cui, Bin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c246t-289374db34b6fa98bb1213b13ea8ebd093a3205ac561a3b2d777728d289e9063</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Data center networks</topic><topic>Data centers</topic><topic>Electronic systems</topic><topic>flowlet</topic><topic>L3 load balancing</topic><topic>Load balancing</topic><topic>Load management</topic><topic>Memory management</topic><topic>Network topology</topic><topic>Protocols</topic><topic>sketch</topic><topic>System-on-chip</topic><topic>Topology</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Liu, Zirui</creatorcontrib><creatorcontrib>Zhao, Yikai</creatorcontrib><creatorcontrib>Fan, Zhuochen</creatorcontrib><creatorcontrib>Yang, Tong</creatorcontrib><creatorcontrib>Li, Xiaodong</creatorcontrib><creatorcontrib>Zhang, Ruwen</creatorcontrib><creatorcontrib>Yang, Kaicheng</creatorcontrib><creatorcontrib>Jiang, Zihan</creatorcontrib><creatorcontrib>Zhong, Zheng</creatorcontrib><creatorcontrib>Huang, Yi</creatorcontrib><creatorcontrib>Liu, Cong</creatorcontrib><creatorcontrib>Hu, Jing</creatorcontrib><creatorcontrib>Xie, Gaogang</creatorcontrib><creatorcontrib>Cui, Bin</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE/IET Electronic Library (IEL)</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Electronics &amp; Communications Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><jtitle>IEEE transactions on parallel and distributed systems</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Liu, Zirui</au><au>Zhao, Yikai</au><au>Fan, Zhuochen</au><au>Yang, Tong</au><au>Li, Xiaodong</au><au>Zhang, Ruwen</au><au>Yang, Kaicheng</au><au>Jiang, Zihan</au><au>Zhong, Zheng</au><au>Huang, Yi</au><au>Liu, Cong</au><au>Hu, Jing</au><au>Xie, Gaogang</au><au>Cui, Bin</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>BurstBalancer: Do Less, Better Balance for Large-Scale Data Center Traffic</atitle><jtitle>IEEE transactions on parallel and distributed systems</jtitle><stitle>TPDS</stitle><date>2024-06-01</date><risdate>2024</risdate><volume>35</volume><issue>6</issue><spage>932</spage><epage>949</epage><pages>932-949</pages><issn>1045-9219</issn><eissn>1558-2183</eissn><coden>ITDSEO</coden><abstract><![CDATA[Layer-3 load balancing is a key topic in the networking field. It is well acknowledged that flowlet is the most promising solution because of its good trade-off between load balance and packet reordering. However, we find its one significant limitation: it makes the forwarding paths of flows unpredictable. To address this limitation, this article presents BurstBalancer, a simple yet efficient load balancing system with a sketch, named BalanceSketch. Our design philosophy is doing less changes to keep the forwarding path of most flows fixed, which guides the design of BalanceSketch and our balance operations. We have fully implemented BurstBalancer in a small-scale testbed built with Tofino switches, and conducted both large-scale event-level (NS-2) and ESL (electronic system level) simulations. Our results show that BurstBalancer achieves 5%<inline-formula><tex-math notation="LaTeX">\sim</tex-math> <mml:math><mml:mo>∼</mml:mo></mml:math><inline-graphic xlink:href="liu-ieq1-3295454.gif"/> </inline-formula>35% smaller FCT than LetFlow in symmetric topology and up to 30× smaller FCT in asymmetric topology, while 58× fewer flows suffer from path changing. All related codes are open-sourced at GitHub.]]></abstract><cop>New York</cop><pub>IEEE</pub><doi>10.1109/TPDS.2023.3295454</doi><tpages>18</tpages><orcidid>https://orcid.org/0000-0003-2402-5854</orcidid><orcidid>https://orcid.org/0000-0002-3534-8045</orcidid><orcidid>https://orcid.org/0000-0003-2495-7774</orcidid><orcidid>https://orcid.org/0000-0003-1681-4677</orcidid><orcidid>https://orcid.org/0000-0002-6102-9195</orcidid><orcidid>https://orcid.org/0000-0001-6381-4026</orcidid><orcidid>https://orcid.org/0000-0003-0042-1828</orcidid><orcidid>https://orcid.org/0000-0002-6741-6751</orcidid><orcidid>https://orcid.org/0000-0003-4964-1135</orcidid><orcidid>https://orcid.org/0000-0003-1160-1058</orcidid><orcidid>https://orcid.org/0000-0001-9062-6565</orcidid></addata></record>
fulltext fulltext
identifier ISSN: 1045-9219
ispartof IEEE transactions on parallel and distributed systems, 2024-06, Vol.35 (6), p.932-949
issn 1045-9219
1558-2183
language eng
recordid cdi_ieee_primary_10184046
source IEEE Xplore (Online service)
subjects Data center networks
Data centers
Electronic systems
flowlet
L3 load balancing
Load balancing
Load management
Memory management
Network topology
Protocols
sketch
System-on-chip
Topology
title BurstBalancer: Do Less, Better Balance for Large-Scale Data Center Traffic
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T21%3A59%3A45IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_ieee_&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=BurstBalancer:%20Do%20Less,%20Better%20Balance%20for%20Large-Scale%20Data%20Center%20Traffic&rft.jtitle=IEEE%20transactions%20on%20parallel%20and%20distributed%20systems&rft.au=Liu,%20Zirui&rft.date=2024-06-01&rft.volume=35&rft.issue=6&rft.spage=932&rft.epage=949&rft.pages=932-949&rft.issn=1045-9219&rft.eissn=1558-2183&rft.coden=ITDSEO&rft_id=info:doi/10.1109/TPDS.2023.3295454&rft_dat=%3Cproquest_ieee_%3E3037644018%3C/proquest_ieee_%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c246t-289374db34b6fa98bb1213b13ea8ebd093a3205ac561a3b2d777728d289e9063%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=3037644018&rft_id=info:pmid/&rft_ieee_id=10184046&rfr_iscdi=true