{"id":"https://openalex.org/W4413756872","doi":"https://doi.org/10.1145/3718958.3750527","title":"SGLB: Scalable and Robust Global Load Balancing in Commodity AI Clusters","display_name":"SGLB: Scalable and Robust Global Load Balancing in Commodity AI Clusters","publication_year":2025,"publication_date":"2025-08-27","ids":{"openalex":"https://openalex.org/W4413756872","doi":"https://doi.org/10.1145/3718958.3750527"},"language":"en","primary_location":{"id":"doi:10.1145/3718958.3750527","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3718958.3750527","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3718958.3750527","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM SIGCOMM 2025 Conference","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3718958.3750527","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5110836348","display_name":"Chengchao Qi","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Chenchen Qi","raw_affiliation_strings":["ByteDance, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0000-2174-6571","affiliations":[{"raw_affiliation_string":"ByteDance, Beijing, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031365162","display_name":"Wenfei Wu","orcid":"https://orcid.org/0000-0002-1357-3137"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wenfei Wu","raw_affiliation_strings":["Peking University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-1357-3137","affiliations":[{"raw_affiliation_string":"Peking University, Beijing, China","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085254179","display_name":"Y.Z. Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yongcan Wang","raw_affiliation_strings":["ByteDance, Hangzhou, China"],"raw_orcid":"https://orcid.org/0009-0001-7215-9753","affiliations":[{"raw_affiliation_string":"ByteDance, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103469967","display_name":"Kan He","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Keqiang He","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0005-7024-9164","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102146037","display_name":"Yu-Hsiang Kao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu-Hsiang (Sean) Kao","raw_affiliation_strings":["ByteDance, San Jose, USA"],"raw_orcid":"https://orcid.org/0009-0004-5424-8340","affiliations":[{"raw_affiliation_string":"ByteDance, San Jose, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006719964","display_name":"Z.D. He","orcid":"https://orcid.org/0009-0007-2249-061X"},"institutions":[{"id":"https://openalex.org/I1296127346","display_name":"Broadcom (Israel)","ror":"https://ror.org/01jsrac29","country_code":"IL","type":"company","lineage":["https://openalex.org/I1296127346","https://openalex.org/I4210127325"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Zongying He","raw_affiliation_strings":["Broadcom, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0007-2249-061X","affiliations":[{"raw_affiliation_string":"Broadcom, Shanghai, China","institution_ids":["https://openalex.org/I1296127346"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006245545","display_name":"Chen-Yu Yen","orcid":"https://orcid.org/0000-0002-5232-5988"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen-Yu Yen","raw_affiliation_strings":["ByteDance, San Jose, USA"],"raw_orcid":"https://orcid.org/0000-0002-5232-5988","affiliations":[{"raw_affiliation_string":"ByteDance, San Jose, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101871699","display_name":"Zhuo Jiang","orcid":"https://orcid.org/0000-0002-6144-7899"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhuo Jiang","raw_affiliation_strings":["ByteDance, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-6144-7899","affiliations":[{"raw_affiliation_string":"ByteDance, Beijing, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100533865","display_name":"Feng Luo","orcid":"https://orcid.org/0009-0000-2012-4658"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Feng Luo","raw_affiliation_strings":["ByteDance, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0000-2012-4658","affiliations":[{"raw_affiliation_string":"ByteDance, Beijing, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5045964353","display_name":"Surendra Anubolu","orcid":null},"institutions":[{"id":"https://openalex.org/I4210127325","display_name":"Broadcom (United States)","ror":"https://ror.org/035gt5s03","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127325"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Surendra Anubolu","raw_affiliation_strings":["Broadcom, San Jose, USA"],"raw_orcid":"https://orcid.org/0009-0005-3450-7445","affiliations":[{"raw_affiliation_string":"Broadcom, San Jose, USA","institution_ids":["https://openalex.org/I4210127325"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069500511","display_name":"Y Gao","orcid":null},"institutions":[{"id":"https://openalex.org/I1296127346","display_name":"Broadcom (Israel)","ror":"https://ror.org/01jsrac29","country_code":"IL","type":"company","lineage":["https://openalex.org/I1296127346","https://openalex.org/I4210127325"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Yanjin Gao","raw_affiliation_strings":["Broadcom, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0008-9758-5741","affiliations":[{"raw_affiliation_string":"Broadcom, Shanghai, China","institution_ids":["https://openalex.org/I1296127346"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Bingfeng Lin","orcid":"https://orcid.org/0009-0009-7807-7175"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bingfeng Lin","raw_affiliation_strings":["ByteDance, Hangzhou, China"],"raw_orcid":"https://orcid.org/0009-0009-7807-7175","affiliations":[{"raw_affiliation_string":"ByteDance, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113857374","display_name":"Wenda Ni","orcid":null},"institutions":[{"id":"https://openalex.org/I58610484","display_name":"Seattle University","ror":"https://ror.org/02jqc0m91","country_code":"US","type":"education","lineage":["https://openalex.org/I58610484"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Wenda Ni","raw_affiliation_strings":["ByteDance, Seattle, USA"],"raw_orcid":"https://orcid.org/0009-0002-7102-5750","affiliations":[{"raw_affiliation_string":"ByteDance, Seattle, USA","institution_ids":["https://openalex.org/I58610484"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yiming Yang","orcid":"https://orcid.org/0009-0006-9216-8039"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yiming Yang","raw_affiliation_strings":["ByteDance, Hangzhou, China"],"raw_orcid":"https://orcid.org/0009-0006-9216-8039","affiliations":[{"raw_affiliation_string":"ByteDance, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Donglin Wei","orcid":"https://orcid.org/0009-0001-7339-6419"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Donglin Wei","raw_affiliation_strings":["ByteDance, Hangzhou, China"],"raw_orcid":"https://orcid.org/0009-0001-7339-6419","affiliations":[{"raw_affiliation_string":"ByteDance, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100646069","display_name":"Boyang Zhou","orcid":"https://orcid.org/0000-0003-0090-2656"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Boyang Zhou","raw_affiliation_strings":["ByteDance, San Jose, USA"],"raw_orcid":"https://orcid.org/0000-0003-0090-2656","affiliations":[{"raw_affiliation_string":"ByteDance, San Jose, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Jian Wang","orcid":"https://orcid.org/0009-0006-3468-5281"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jian Wang","raw_affiliation_strings":["ByteDance, Hangzhou, China"],"raw_orcid":"https://orcid.org/0009-0006-3468-5281","affiliations":[{"raw_affiliation_string":"ByteDance, Hangzhou, China","institution_ids":[]}]},{"author_position":"last","author":{"id":null,"display_name":"Shan Ding","orcid":"https://orcid.org/0009-0007-1875-3085"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shan Ding","raw_affiliation_strings":["ByteDance, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0007-1875-3085","affiliations":[{"raw_affiliation_string":"ByteDance, Beijing, China","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":18,"corresponding_author_ids":["https://openalex.org/A5110836348"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":5.7202,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.96156871,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"626","last_page":"644"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10101","display_name":"Cloud Computing and Resource Management","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10101","display_name":"Cloud Computing and Resource Management","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.9993000030517578,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12292","display_name":"Graph Theory and Algorithms","score":0.9965000152587891,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.7285033464431763},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7062262296676636},{"id":"https://openalex.org/keywords/load-balancing","display_name":"Load balancing (electrical power)","score":0.6438949704170227},{"id":"https://openalex.org/keywords/commodity","display_name":"Commodity","score":0.5853305459022522},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.4538801908493042},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.16797369718551636},{"id":"https://openalex.org/keywords/economics","display_name":"Economics","score":0.1108052134513855},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.0604514479637146}],"concepts":[{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7285033464431763},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7062262296676636},{"id":"https://openalex.org/C138959212","wikidata":"https://www.wikidata.org/wiki/Q1806783","display_name":"Load balancing (electrical power)","level":3,"score":0.6438949704170227},{"id":"https://openalex.org/C2779439359","wikidata":"https://www.wikidata.org/wiki/Q317088","display_name":"Commodity","level":2,"score":0.5853305459022522},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.4538801908493042},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.16797369718551636},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.1108052134513855},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0604514479637146},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0},{"id":"https://openalex.org/C34447519","wikidata":"https://www.wikidata.org/wiki/Q179522","display_name":"Market economy","level":1,"score":0.0},{"id":"https://openalex.org/C187691185","wikidata":"https://www.wikidata.org/wiki/Q2020720","display_name":"Grid","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3718958.3750527","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3718958.3750527","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3718958.3750527","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM SIGCOMM 2025 Conference","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3718958.3750527","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3718958.3750527","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3718958.3750527","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM SIGCOMM 2025 Conference","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.46000000834465027,"display_name":"Decent work and economic growth","id":"https://metadata.un.org/sdg/8"}],"awards":[{"id":"https://openalex.org/G3861296800","display_name":null,"funder_award_id":"62472009","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4413756872.pdf","grobid_xml":"https://content.openalex.org/works/W4413756872.grobid-xml"},"referenced_works_count":29,"referenced_works":["https://openalex.org/W95608104","https://openalex.org/W2087178199","https://openalex.org/W2126969025","https://openalex.org/W2132320636","https://openalex.org/W2143163492","https://openalex.org/W2157614013","https://openalex.org/W2157990152","https://openalex.org/W2169246522","https://openalex.org/W2465793152","https://openalex.org/W2498764059","https://openalex.org/W2744387122","https://openalex.org/W2744567246","https://openalex.org/W2744698795","https://openalex.org/W2884336345","https://openalex.org/W2913574348","https://openalex.org/W2969388332","https://openalex.org/W3129251660","https://openalex.org/W3188706325","https://openalex.org/W3197720002","https://openalex.org/W4205118271","https://openalex.org/W4230033334","https://openalex.org/W4235670058","https://openalex.org/W4235732943","https://openalex.org/W4243367687","https://openalex.org/W4249923717","https://openalex.org/W4300490703","https://openalex.org/W4386385105","https://openalex.org/W4401176521","https://openalex.org/W4401176799"],"related_works":["https://openalex.org/W2166734474","https://openalex.org/W1982914007","https://openalex.org/W3004117068","https://openalex.org/W2043284556","https://openalex.org/W2159583675","https://openalex.org/W3175372827","https://openalex.org/W2488220206","https://openalex.org/W2146301767","https://openalex.org/W1824242903","https://openalex.org/W1493858311"],"abstract_inverted_index":{"Internet":[0],"companies":[1],"are":[2,133],"constructing":[3],"large-scale":[4],"AI":[5,12,19,56],"clusters":[6],"with":[7,181],"commodity":[8,73],"Ethernet":[9],"switches":[10],"for":[11,55,167],"model":[13,170],"training":[14,20],"to":[15,64,83,101,142,163,191],"support":[16],"their":[17],"businesses.":[18],"workloads":[21],"impose":[22],"stringent":[23],"network":[24,165],"requirements,":[25],"mandating":[26],"that":[27],"cluster":[28],"networks":[29],"deliver":[30],"high":[31],"peak":[32],"throughput":[33,126],"while":[34],"maintaining":[35],"robustness":[36,166],"and":[37,123,138],"resilience":[38],"in":[39,71,95,104,128,150,157],"the":[40,105,116,151],"face":[41],"of":[42,107,112,118,153],"link":[43,108,154],"failures.":[44],"We":[45,90,135],"present":[46],"SGLB,":[47],"a":[48,60,66],"distributed,":[49],"global":[50,80,182],"congestion-aware":[51],"load":[52,68],"balancing":[53,69],"system":[54],"clusters.":[57],"SGLB":[58,137,145,173],"operates":[59],"control-plane":[61],"protocol,":[62],"SyncMesh,":[63],"enable":[65],"new":[67],"abstraction":[70],"modern":[72],"switches\u2014Global":[74],"Load":[75],"Balancing":[76],"(GLB)":[77],"engine\u2014which":[78],"utilizes":[79],"congestion":[81,113],"information":[82],"distribute":[84],"traffic":[85,176],"across":[86,177],"all":[87],"available":[88],"paths.":[89],"address":[91],"three":[92],"key":[93],"challenges":[94],"designing":[96],"SGLB:":[97],"fast":[98],"routing":[99,148],"convergence":[100,149],"minimize":[102],"downtime":[103],"event":[106,152],"failures,":[109,155],"scalable":[110],"maintenance":[111],"profiles":[114],"within":[115],"constraints":[117],"limited":[119],"switch":[120],"hardware":[121],"resources,":[122],"preventing":[124],"GLB":[125],"suppression":[127],"scenarios":[129],"where":[130],"path":[131],"bandwidths":[132],"asymmetric.":[134],"prototype":[136],"conduct":[139],"extensive":[140],"experiments":[141],"evaluate":[143],"SGLB.":[144],"ensures":[146],"rapid":[147],"recovering":[156],"as":[158,160],"little":[159],"45":[161],"\u03bcs":[162],"guarantee":[164],"long-term,":[168],"stable":[169],"training.":[171],"Additionally,":[172],"effectively":[174],"load-balances":[175],"paths,":[178],"avoiding":[179],"those":[180],"congestion,":[183],"which":[184],"accelerates":[185],"All-to-All":[186],"collective":[187],"communication":[188],"by":[189],"up":[190],"60%.":[192]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-04-30T09:15:22.047038","created_date":"2025-10-10T00:00:00"}
