{"id":"https://openalex.org/W4400682270","doi":"https://doi.org/10.1145/3672198.3673800","title":"In-Network AllReduce Optimization with Virtual Aggregation Trees","display_name":"In-Network AllReduce Optimization with Virtual Aggregation Trees","publication_year":2024,"publication_date":"2024-07-16","ids":{"openalex":"https://openalex.org/W4400682270","doi":"https://doi.org/10.1145/3672198.3673800"},"language":"en","primary_location":{"id":"doi:10.1145/3672198.3673800","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3672198.3673800","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2024 SIGCOMM Workshop on Networks for AI Computing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5026896241","display_name":"Haoyu Song","orcid":"https://orcid.org/0000-0001-5377-6628"},"institutions":[{"id":"https://openalex.org/I4210146936","display_name":"Huawei Technologies (United States)","ror":"https://ror.org/03jyqk712","country_code":"US","type":"company","lineage":["https://openalex.org/I2250955327","https://openalex.org/I4210146936"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Haoyu Song","raw_affiliation_strings":["Futurewei Technologies, Santa Clara, CA, USA"],"raw_orcid":"https://orcid.org/0000-0001-5377-6628","affiliations":[{"raw_affiliation_string":"Futurewei Technologies, Santa Clara, CA, USA","institution_ids":["https://openalex.org/I4210146936"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5026896241"],"corresponding_institution_ids":["https://openalex.org/I4210146936"],"apc_list":null,"apc_paid":null,"fwci":0.6728,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.7019287,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"54","last_page":"60"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10714","display_name":"Software-Defined Networks and 5G","score":0.9988999962806702,"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"}},"topics":[{"id":"https://openalex.org/T10714","display_name":"Software-Defined Networks and 5G","score":0.9988999962806702,"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/T10101","display_name":"Cloud Computing and Resource Management","score":0.9980000257492065,"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/T10829","display_name":"Interconnection Networks and Systems","score":0.9975000023841858,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6476550102233887}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6476550102233887}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3672198.3673800","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3672198.3673800","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2024 SIGCOMM Workshop on Networks for AI Computing","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":11,"referenced_works":["https://openalex.org/W95608104","https://openalex.org/W2072811945","https://openalex.org/W2292566746","https://openalex.org/W2494561778","https://openalex.org/W2761338514","https://openalex.org/W2769986458","https://openalex.org/W2789851468","https://openalex.org/W4242634435","https://openalex.org/W4290991009","https://openalex.org/W4318540560","https://openalex.org/W4327911839"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"AllReduce":[0,125],"is":[1],"a":[2,66,123,134],"critical":[3],"performance":[4],"bottleneck":[5],"for":[6,17,45,117,137],"distributed":[7],"deep":[8],"learning":[9],"and":[10,57,91],"large":[11],"model":[12],"training":[13],"in":[14,152],"data":[15,80],"centers":[16],"AI":[18],"computing.":[19],"In-Network":[20],"Aggregation":[21,72],"(INA)":[22],"has":[23],"been":[24],"identified":[25],"as":[26],"an":[27],"effective":[28],"accelerating":[29],"technique":[30],"to":[31,69,84],"improve":[32],"its":[33],"performance.":[34],"However,":[35],"the":[36,42,55,78,105,113,147,155,164],"existing":[37],"schemes":[38,151],"either":[39],"use":[40],"only":[41],"ToR":[43],"switches":[44,90],"aggregation":[46,51,118],"or":[47,166],"rely":[48],"on":[49,75,161],"physical":[50,79],"trees,":[52],"which":[53],"limit":[54],"efficiency":[56],"scalablility":[58],"of":[59,77,87,154,157],"INA.":[60],"In":[61],"this":[62],"paper":[63],"we":[64,121],"propose":[65],"novel":[67],"method":[68],"construct":[70],"Virtual":[71],"Trees":[73],"(VAT)":[74],"top":[76],"center":[81],"network":[82,89],"topology":[83],"take":[85],"advantage":[86],"all":[88],"balance":[92],"their":[93],"load.":[94],"We":[95,132],"show":[96,143],"that":[97,144],"VAT":[98,139,145],"encoding":[99],"can":[100],"be":[101],"realized":[102],"by":[103,110],"repurposing":[104],"IETF":[106],"BIER":[107,114],"bitmap.":[108],"Thus,":[109],"integrating":[111],"with":[112,127,163],"multicast":[115],"feature":[116],"result":[119],"distribution,":[120],"have":[122],"complete":[124],"solution":[126],"network-facilitated":[128],"collective":[129],"communication":[130],"optimizations.":[131],"develop":[133],"greedy":[135],"algorithm":[136],"incremental":[138],"construction.":[140],"The":[141],"evaluations":[142],"outperforms":[146],"other":[148],"state-of-art":[149],"INA":[150],"terms":[153],"number":[156],"parallel":[158],"jobs":[159],"supported":[160],"networks":[162],"Spine-Leaf":[165],"Fat-tree":[167],"topologies.":[168]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2025-12-21T01:58:51.020947","created_date":"2025-10-10T00:00:00"}