{"id":"https://openalex.org/W4294982491","doi":"https://doi.org/10.1109/tnet.2022.3202529","title":"Deep Learning-Based Job Placement in Distributed Machine Learning Clusters With Heterogeneous Workloads","display_name":"Deep Learning-Based Job Placement in Distributed Machine Learning Clusters With Heterogeneous Workloads","publication_year":2022,"publication_date":"2022-09-08","ids":{"openalex":"https://openalex.org/W4294982491","doi":"https://doi.org/10.1109/tnet.2022.3202529"},"language":"en","primary_location":{"id":"doi:10.1109/tnet.2022.3202529","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tnet.2022.3202529","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/A5054679083","display_name":"Yixin Bao","orcid":"https://orcid.org/0000-0002-6921-2154"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yixin Bao","raw_affiliation_strings":["Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong"],"raw_orcid":"https://orcid.org/0000-0002-6921-2154","affiliations":[{"raw_affiliation_string":"Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013856812","display_name":"Yanghua Peng","orcid":"https://orcid.org/0000-0003-3989-4358"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yanghua Peng","raw_affiliation_strings":["Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong"],"raw_orcid":"https://orcid.org/0000-0003-3989-4358","affiliations":[{"raw_affiliation_string":"Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5012597518","display_name":"Chuan Wu","orcid":"https://orcid.org/0000-0002-3144-4398"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Chuan Wu","raw_affiliation_strings":["Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong"],"raw_orcid":"https://orcid.org/0000-0002-3144-4398","affiliations":[{"raw_affiliation_string":"Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I889458895"],"apc_list":null,"apc_paid":null,"fwci":8.7224,"has_fulltext":false,"cited_by_count":31,"citation_normalized_percentile":{"value":0.97728438,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":"31","issue":"2","first_page":"634","last_page":"647"},"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/T10273","display_name":"IoT and Edge/Fog Computing","score":0.9984999895095825,"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/T13553","display_name":"Age of Information Optimization","score":0.9944999814033508,"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":0.8770397901535034},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.6562567353248596},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.623813807964325},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.5306001901626587},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.4995396137237549},{"id":"https://openalex.org/keywords/server","display_name":"Server","score":0.49557673931121826},{"id":"https://openalex.org/keywords/workload","display_name":"Workload","score":0.4657253921031952},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.4636920690536499},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.18512415885925293},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.15296441316604614}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8770397901535034},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6562567353248596},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.623813807964325},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5306001901626587},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4995396137237549},{"id":"https://openalex.org/C93996380","wikidata":"https://www.wikidata.org/wiki/Q44127","display_name":"Server","level":2,"score":0.49557673931121826},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.4657253921031952},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.4636920690536499},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.18512415885925293},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.15296441316604614}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tnet.2022.3202529","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tnet.2022.3202529","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":[{"display_name":"Decent work and economic growth","id":"https://metadata.un.org/sdg/8","score":0.6600000262260437}],"awards":[{"id":"https://openalex.org/G5807435805","display_name":"Accelerating Distributed DNN Training through Fine-grained Communication and Computation Placement and Scheduling","funder_award_id":"17208920","funder_id":"https://openalex.org/F4320321592","funder_display_name":"Research Grants Council, University Grants Committee"},{"id":"https://openalex.org/G692013346","display_name":null,"funder_award_id":"HKU 17204619","funder_id":"https://openalex.org/F4320321592","funder_display_name":"Research Grants Council, University Grants Committee"},{"id":"https://openalex.org/G7753977827","display_name":"Optimizing Distributed GNN Training on Large Graphs","funder_award_id":"17207621","funder_id":"https://openalex.org/F4320321592","funder_display_name":"Research Grants Council, University Grants Committee"}],"funders":[{"id":"https://openalex.org/F4320321592","display_name":"Research Grants Council, University Grants Committee","ror":"https://ror.org/00djwmt25"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":82,"referenced_works":["https://openalex.org/W1442374986","https://openalex.org/W1522301498","https://openalex.org/W1614298861","https://openalex.org/W1665214252","https://openalex.org/W1686810756","https://openalex.org/W1832693441","https://openalex.org/W2021667890","https://openalex.org/W2057332538","https://openalex.org/W2060393849","https://openalex.org/W2064675550","https://openalex.org/W2103946820","https://openalex.org/W2105947650","https://openalex.org/W2129542763","https://openalex.org/W2130942839","https://openalex.org/W2133569115","https://openalex.org/W2139258336","https://openalex.org/W2141992894","https://openalex.org/W2155027007","https://openalex.org/W2163605009","https://openalex.org/W2163961697","https://openalex.org/W2186615578","https://openalex.org/W2189062930","https://openalex.org/W2194775991","https://openalex.org/W2461411193","https://openalex.org/W2546571074","https://openalex.org/W2597068831","https://openalex.org/W2609354095","https://openalex.org/W2617411258","https://openalex.org/W2744628735","https://openalex.org/W2749807327","https://openalex.org/W2766447205","https://openalex.org/W2786414509","https://openalex.org/W2798515322","https://openalex.org/W2899071864","https://openalex.org/W2918828872","https://openalex.org/W2919594608","https://openalex.org/W2949562459","https://openalex.org/W2953384591","https://openalex.org/W2962684017","https://openalex.org/W2962725887","https://openalex.org/W2963549123","https://openalex.org/W2964043796","https://openalex.org/W2968986602","https://openalex.org/W3022298203","https://openalex.org/W3037519745","https://openalex.org/W3096583839","https://openalex.org/W3097875570","https://openalex.org/W3102385572","https://openalex.org/W3121689374","https://openalex.org/W4214717370","https://openalex.org/W4235985402","https://openalex.org/W4236491543","https://openalex.org/W4253824360","https://openalex.org/W4288289123","https://openalex.org/W4295312788","https://openalex.org/W4301239768","https://openalex.org/W4385245566","https://openalex.org/W6628377381","https://openalex.org/W6631190155","https://openalex.org/W6636510571","https://openalex.org/W6637242042","https://openalex.org/W6637373629","https://openalex.org/W6671887237","https://openalex.org/W6679436768","https://openalex.org/W6683204974","https://openalex.org/W6684084819","https://openalex.org/W6684191040","https://openalex.org/W6686509673","https://openalex.org/W6687460195","https://openalex.org/W6692846177","https://openalex.org/W6713134421","https://openalex.org/W6738144653","https://openalex.org/W6739901393","https://openalex.org/W6743756900","https://openalex.org/W6748645090","https://openalex.org/W6748687944","https://openalex.org/W6756009870","https://openalex.org/W6759814162","https://openalex.org/W6766978945","https://openalex.org/W6769424276","https://openalex.org/W6779740189","https://openalex.org/W6782839094"],"related_works":["https://openalex.org/W986318368","https://openalex.org/W2000785801","https://openalex.org/W1982914007","https://openalex.org/W2159583675","https://openalex.org/W1852113167","https://openalex.org/W1493858311","https://openalex.org/W1824242903","https://openalex.org/W2155470929","https://openalex.org/W2394465510","https://openalex.org/W2111125783"],"abstract_inverted_index":{"Nowadays,":[0],"most":[1],"leading":[2,69],"IT":[3],"companies":[4],"host":[5],"a":[6,37,96,105,125,145,160,189,228],"variety":[7],"of":[8,140,188,192,231,245],"distributed":[9],"machine":[10,25],"learning":[11,150],"(ML)":[12],"workloads":[13,226],"in":[14,36,65,124,227,243],"ML":[15,47,108,225],"clusters":[16],"to":[17,40,51,70,168,196],"support":[18],"AI-driven":[19],"services,":[20],"such":[21,58],"as":[22,59],"speech":[23],"recognition,":[24],"translation,":[26],"and":[27,61,74,90,130,171,183,214],"image":[28],"processing.":[29],"While":[30],"multiple":[31],"jobs":[32,48,115],"are":[33,63],"executed":[34],"concurrently":[35],"shared":[38],"cluster":[39,56,109,230],"improve":[41,172],"resource":[42,72],"utilization,":[43],"interference":[44,91,129],"among":[45],"co-located":[46],"can":[49],"lead":[50],"significant":[52],"performance":[53,132],"downgrade.":[54],"Existing":[55],"schedulers,":[57],"YARN":[60],"Mesos,":[62],"interference-agnostic":[64],"their":[66],"job":[67,81,247],"placement,":[68],"suboptimal":[71],"efficiency":[73],"usage.":[75],"Some":[76],"literature":[77],"has":[78],"studied":[79],"interference-aware":[80],"placement":[82,198],"policy,":[83],"but":[84],"relies":[85],"on":[86,144],"detailed":[87],"workload":[88],"profiling":[89],"modeling,":[92],"which":[93,208],"is":[94,142,209],"not":[95],"general":[97],"solution.":[98],"In":[99,186],"this":[100],"work,":[101],"we":[102,200],"present":[103],"Harmony,":[104],"deep":[106,148],"learning-driven":[107],"scheduler":[110],"that":[111,127,236],"places":[112],"heterogeneous":[113],"training":[114,134,170],"(either":[116],"with":[117,154,164,211],"parameter":[118],"server":[119],"architecture":[120],"or":[121],"all-reduce":[122],"architecture)":[123],"manner":[126],"minimizes":[128],"maximizes":[131],"(i.e.,":[133],"completion":[135,248],"time":[136],"minimization).":[137],"The":[138,157],"design":[139],"Harmony":[141,237],"based":[143],"carefully":[146],"designed":[147],"reinforcement":[149],"(DRL)":[151],"framework":[152],"enhanced":[153],"reward":[155,193,205,218],"modeling.":[156],"DRL":[158],"integrates":[159],"dynamic":[161],"sequence-to-sequence":[162,204],"model":[163],"the":[165],"state-of-the-art":[166],"techniques":[167],"stabilize":[169],"convergence,":[173],"including":[174],"actor-critic":[175],"algorithm,":[176],"job-aware":[177],"action":[178],"space":[179],"exploration,":[180],"multi-head":[181],"attention,":[182],"experience":[184],"replay.":[185],"view":[187],"common":[190],"lack":[191],"samples":[194,213],"corresponding":[195],"different":[197],"decisions,":[199],"build":[201],"an":[202],"auxiliary":[203],"prediction":[206],"model,":[207],"trained":[210],"historical":[212],"used":[215],"for":[216,219],"producing":[217],"unseen":[220],"placement.":[221],"Experiments":[222],"using":[223],"real":[224],"Kubernetes":[229],"6":[232],"GPU":[233],"servers":[234],"show":[235],"outperforms":[238],"representative":[239],"schedulers":[240],"by":[241],"16%\u201342%":[242],"terms":[244],"average":[246],"time.":[249]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":16},{"year":2024,"cited_by_count":6},{"year":2023,"cited_by_count":6}],"updated_date":"2026-07-22T07:51:19.307946","created_date":"2025-10-10T00:00:00"}
