{"id":"https://openalex.org/W4389722097","doi":"https://doi.org/10.1109/tnet.2023.3339524","title":"XAgg: Accelerating Heterogeneous Distributed Training Through XDP-Based Gradient Aggregation","display_name":"XAgg: Accelerating Heterogeneous Distributed Training Through XDP-Based Gradient Aggregation","publication_year":2023,"publication_date":"2023-12-14","ids":{"openalex":"https://openalex.org/W4389722097","doi":"https://doi.org/10.1109/tnet.2023.3339524"},"language":"en","primary_location":{"id":"doi:10.1109/tnet.2023.3339524","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tnet.2023.3339524","pdf_url":null,"source":{"id":"https://openalex.org/S62238642","display_name":"IEEE/ACM Transactions on Networking","issn_l":"1063-6692","issn":["1063-6692","1558-2566"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE/ACM Transactions on Networking","raw_type":"journal-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/A5101466146","display_name":"Qianyu Zhang","orcid":"https://orcid.org/0000-0002-3805-4480"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Qianyu Zhang","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China","Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I126520041"]},{"raw_affiliation_string":"Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074187380","display_name":"Gongming Zhao","orcid":"https://orcid.org/0000-0003-1311-8908"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Gongming Zhao","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China","Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I126520041"]},{"raw_affiliation_string":"Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5063184427","display_name":"Hongli Xu","orcid":"https://orcid.org/0000-0003-3831-4577"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hongli Xu","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China","Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I126520041"]},{"raw_affiliation_string":"Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5115621130","display_name":"Peng Yang","orcid":"https://orcid.org/0000-0003-4400-2289"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Peng Yang","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China","Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I126520041"]},{"raw_affiliation_string":"Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5101466146"],"corresponding_institution_ids":["https://openalex.org/I126520041"],"apc_list":null,"apc_paid":null,"fwci":0.8405,"has_fulltext":false,"cited_by_count":7,"citation_normalized_percentile":{"value":0.76029865,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":96,"max":99},"biblio":{"volume":"32","issue":"3","first_page":"2174","last_page":"2188"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9991999864578247,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9991999864578247,"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"}},{"id":"https://openalex.org/T10101","display_name":"Cloud Computing and Resource Management","score":0.9957000017166138,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9957000017166138,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8082683086395264},{"id":"https://openalex.org/keywords/asynchronous-communication","display_name":"Asynchronous communication","score":0.6199010014533997},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.539070188999176},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.4970252811908722},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.46710529923439026},{"id":"https://openalex.org/keywords/transmission","display_name":"Transmission (telecommunications)","score":0.4622260332107544},{"id":"https://openalex.org/keywords/synchronization","display_name":"Synchronization (alternating current)","score":0.44807130098342896},{"id":"https://openalex.org/keywords/server","display_name":"Server","score":0.4352019131183624},{"id":"https://openalex.org/keywords/testbed","display_name":"Testbed","score":0.4316396415233612},{"id":"https://openalex.org/keywords/real-time-computing","display_name":"Real-time computing","score":0.33021658658981323},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.3282511830329895},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3251349925994873},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.2558076083660126},{"id":"https://openalex.org/keywords/channel","display_name":"Channel (broadcasting)","score":0.1215391755104065}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8082683086395264},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.6199010014533997},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.539070188999176},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.4970252811908722},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.46710529923439026},{"id":"https://openalex.org/C761482","wikidata":"https://www.wikidata.org/wiki/Q118093","display_name":"Transmission (telecommunications)","level":2,"score":0.4622260332107544},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.44807130098342896},{"id":"https://openalex.org/C93996380","wikidata":"https://www.wikidata.org/wiki/Q44127","display_name":"Server","level":2,"score":0.4352019131183624},{"id":"https://openalex.org/C31395832","wikidata":"https://www.wikidata.org/wiki/Q1318674","display_name":"Testbed","level":2,"score":0.4316396415233612},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.33021658658981323},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.3282511830329895},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3251349925994873},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.2558076083660126},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.1215391755104065},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tnet.2023.3339524","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tnet.2023.3339524","pdf_url":null,"source":{"id":"https://openalex.org/S62238642","display_name":"IEEE/ACM Transactions on Networking","issn_l":"1063-6692","issn":["1063-6692","1558-2566"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE/ACM Transactions on Networking","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1458367098","display_name":null,"funder_award_id":"62372426","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G1745555229","display_name":null,"funder_award_id":"2023481","funder_id":"https://openalex.org/F4320321133","funder_display_name":"Chinese Academy of Sciences"},{"id":"https://openalex.org/G7563642939","display_name":null,"funder_award_id":"BK20210121","funder_id":"https://openalex.org/F4320322769","funder_display_name":"Natural Science Foundation of Jiangsu Province"},{"id":"https://openalex.org/G7748147380","display_name":null,"funder_award_id":"62102392","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"},{"id":"https://openalex.org/F4320321133","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35"},{"id":"https://openalex.org/F4320322769","display_name":"Natural Science Foundation of Jiangsu Province","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320335787","display_name":"Fundamental Research Funds for the Central Universities","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":72,"referenced_works":["https://openalex.org/W1686810756","https://openalex.org/W1966759369","https://openalex.org/W1991055327","https://openalex.org/W2034673442","https://openalex.org/W2060393849","https://openalex.org/W2099338166","https://openalex.org/W2113487548","https://openalex.org/W2117539524","https://openalex.org/W2129807746","https://openalex.org/W2143295630","https://openalex.org/W2183341477","https://openalex.org/W2186615578","https://openalex.org/W2194775991","https://openalex.org/W2402144811","https://openalex.org/W2524428287","https://openalex.org/W2606278415","https://openalex.org/W2769986458","https://openalex.org/W2774000609","https://openalex.org/W2884711234","https://openalex.org/W2887718373","https://openalex.org/W2896457183","https://openalex.org/W2903038868","https://openalex.org/W2906731962","https://openalex.org/W2912174032","https://openalex.org/W2948952649","https://openalex.org/W2953169926","https://openalex.org/W2962843773","https://openalex.org/W2973727699","https://openalex.org/W2975712713","https://openalex.org/W2991040477","https://openalex.org/W2998811572","https://openalex.org/W3002226419","https://openalex.org/W3011196476","https://openalex.org/W3011594683","https://openalex.org/W3012220622","https://openalex.org/W3047002910","https://openalex.org/W3091097978","https://openalex.org/W3093671646","https://openalex.org/W3134069806","https://openalex.org/W3155963862","https://openalex.org/W3192241394","https://openalex.org/W4236757187","https://openalex.org/W4254760511","https://openalex.org/W4288083766","https://openalex.org/W4289300273","https://openalex.org/W4289401659","https://openalex.org/W4301239768","https://openalex.org/W4385245566","https://openalex.org/W6636548729","https://openalex.org/W6637373629","https://openalex.org/W6684859321","https://openalex.org/W6686509673","https://openalex.org/W6715382103","https://openalex.org/W6727208969","https://openalex.org/W6738460352","https://openalex.org/W6746839373","https://openalex.org/W6748645090","https://openalex.org/W6750584272","https://openalex.org/W6753209298","https://openalex.org/W6754818105","https://openalex.org/W6755207826","https://openalex.org/W6756592394","https://openalex.org/W6757449384","https://openalex.org/W6760703128","https://openalex.org/W6763300512","https://openalex.org/W6766978945","https://openalex.org/W6767997687","https://openalex.org/W6772544576","https://openalex.org/W6775649616","https://openalex.org/W6784425352","https://openalex.org/W6794216301","https://openalex.org/W6794224617"],"related_works":["https://openalex.org/W2883256816","https://openalex.org/W2171408034","https://openalex.org/W3003320923","https://openalex.org/W2106140982","https://openalex.org/W2595172197","https://openalex.org/W2152313554","https://openalex.org/W2064303750","https://openalex.org/W4285042611","https://openalex.org/W1509300825","https://openalex.org/W3092582874"],"abstract_inverted_index":{"With":[0],"the":[1,12,46,64,79,83,90,95,108,118,126,167],"growth":[2],"of":[3,23,50],"model/dataset/system":[4],"size":[5],"for":[6],"distributed":[7,65],"model":[8],"training":[9,66],"in":[10,71,138],"datacenters,":[11],"widely":[13],"used":[14],"Parameter":[15],"Server":[16],"(PS)":[17],"architecture":[18],"suffers":[19],"from":[20],"communication":[21,40,194],"bottleneck":[22],"gradient":[24,36,53,80,104,136,148,168],"transmission.":[25],"Recent":[26],"works":[27],"attempt":[28],"to":[29,33,45,58,131],"utilize":[30],"programmable":[31,51],"switches":[32],"implement":[34],"in-network":[35],"aggregation":[37,61,84,91,105,169],"and":[38,75,78,146,153,159],"alleviate":[39],"bottlenecks":[41],"on":[42,115,122],"PSs.":[43],"Due":[44],"limited":[47],"on-chip":[48],"memory":[49,121],"switches,":[52],"transmission":[54,137],"requires":[55],"strict":[56],"synchronization":[57],"achieve":[59],"ideal":[60],"performance.":[62,92],"However,":[63],"system":[67],"is":[68],"usually":[69],"heterogeneous":[70,103,139],"datacenters":[72],"(e.g.,":[73],"computation":[74],"bandwidth":[76],"heterogeneity),":[77],"will":[81],"reach":[82],"nodes":[85],"asynchronously,":[86],"thereby":[87],"seriously":[88],"affecting":[89],"To":[93],"solve":[94],"above":[96],"issue,":[97],"we":[98],"propose":[99],"XAgg,":[100],"which":[101],"accelerates":[102],"by":[106,171,196],"deploying":[107],"eXpress":[109],"Data":[110],"Path":[111],"(XDP)":[112],"based":[113],"aggregator":[114],"servers.":[116],"Specifically,":[117],"abundant":[119],"idle":[120],"servers":[123],"can":[124,143],"cache":[125],"entire":[127],"gradient,":[128],"so":[129],"as":[130],"effectively":[132],"deal":[133],"with":[134,155,180,186,199],"asynchronous":[135],"scenarios.":[140],"Moreover,":[141],"XDP":[142],"provide":[144],"high-performance":[145],"low-latency":[147],"aggregation.":[149],"We":[150],"conduct":[151],"microbenchmark":[152],"testbed":[154],"real-world":[156],"DNN":[157],"models":[158],"datasets.":[160],"Experimental":[161],"results":[162],"show":[163],"that":[164],"XAgg":[165,192],"improves":[166],"throughput":[170],"3.3":[172],"<inline-formula":[173],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[174],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">":[175],"<tex-math":[176],"notation=\"LaTeX\">$\\times$</tex-math>":[177],"</inline-formula>":[178],"compared":[179,198],"TCP-based":[181],"aggregation,":[182],"reaching":[183],"100":[184],"Gbps":[185],"10":[187],"CPU":[188],"cores.":[189],"In":[190],"addition,":[191],"reduces":[193],"time":[195],"49%-82%":[197],"state-of-the-art":[200],"solutions.":[201]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":3},{"year":2024,"cited_by_count":3}],"updated_date":"2026-04-02T15:55:50.835912","created_date":"2025-10-10T00:00:00"}
