{"id":"https://openalex.org/W4415748021","doi":"https://doi.org/10.1109/tpds.2025.3627574","title":"Joint Optimization of Resource Allocation and Request Batching for Multi-Tenant Inference Serving on GPU","display_name":"Joint Optimization of Resource Allocation and Request Batching for Multi-Tenant Inference Serving on GPU","publication_year":2025,"publication_date":"2025-10-31","ids":{"openalex":"https://openalex.org/W4415748021","doi":"https://doi.org/10.1109/tpds.2025.3627574"},"language":null,"primary_location":{"id":"doi:10.1109/tpds.2025.3627574","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2025.3627574","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"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 Transactions on Parallel and Distributed Systems","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/A5039364773","display_name":"Yuning Zhang","orcid":"https://orcid.org/0000-0002-0306-1859"},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"The University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":true,"raw_author_name":"Yuning Zhang","raw_affiliation_strings":["School of Electrical and Computer Engineering, University of Sydney, Darlington, NSW, Australia","School of Electrical and Computer Engineering, The University of Sydney, Darlington, NSW, Australia"],"raw_orcid":"https://orcid.org/0000-0002-0306-1859","affiliations":[{"raw_affiliation_string":"School of Electrical and Computer Engineering, University of Sydney, Darlington, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]},{"raw_affiliation_string":"School of Electrical and Computer Engineering, The University of Sydney, Darlington, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Nan Yang","orcid":"https://orcid.org/0000-0001-5257-9227"},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"The University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Nan Yang","raw_affiliation_strings":["School of Electrical and Computer Engineering, University of Sydney, Darlington, NSW, Australia","School of Electrical and Computer Engineering, The University of Sydney, Darlington, NSW, Australia"],"raw_orcid":"https://orcid.org/0000-0001-5257-9227","affiliations":[{"raw_affiliation_string":"School of Electrical and Computer Engineering, University of Sydney, Darlington, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]},{"raw_affiliation_string":"School of Electrical and Computer Engineering, The University of Sydney, Darlington, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Chen Pan","orcid":"https://orcid.org/0009-0008-7891-4474"},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"The University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Chen Pan","raw_affiliation_strings":["School of Electrical and Computer Engineering, University of Sydney, Darlington, NSW, Australia","School of Electrical and Computer Engineering, The University of Sydney, Darlington, NSW, Australia"],"raw_orcid":"https://orcid.org/0009-0008-7891-4474","affiliations":[{"raw_affiliation_string":"School of Electrical and Computer Engineering, University of Sydney, Darlington, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]},{"raw_affiliation_string":"School of Electrical and Computer Engineering, The University of Sydney, Darlington, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5054168288","display_name":"Dong Yuan","orcid":null},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"The University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Dong Yuan","raw_affiliation_strings":["School of Electrical and Computer Engineering, University of Sydney, Darlington, NSW, Australia","School of Electrical and Computer Engineering, The University of Sydney, Darlington, NSW, Australia"],"raw_orcid":"https://orcid.org/0000-0003-1130-0888","affiliations":[{"raw_affiliation_string":"School of Electrical and Computer Engineering, University of Sydney, Darlington, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]},{"raw_affiliation_string":"School of Electrical and Computer Engineering, The University of Sydney, Darlington, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5039364773"],"corresponding_institution_ids":["https://openalex.org/I129604602"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.33470437,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"37","issue":"1","first_page":"287","last_page":"303"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.3375999927520752,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.3375999927520752,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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":0.19460000097751617,"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/T14347","display_name":"Big Data and Digital Economy","score":0.16009999811649323,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.7760000228881836},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.758899986743927},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.6603999733924866},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.5223000049591064},{"id":"https://openalex.org/keywords/resource-allocation","display_name":"Resource allocation","score":0.46219998598098755},{"id":"https://openalex.org/keywords/job-scheduler","display_name":"Job scheduler","score":0.45320001244544983},{"id":"https://openalex.org/keywords/job-shop-scheduling","display_name":"Job shop scheduling","score":0.4196000099182129},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.400299996137619}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.90829998254776},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.7760000228881836},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.758899986743927},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6603999733924866},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.5223000049591064},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.51910001039505},{"id":"https://openalex.org/C29202148","wikidata":"https://www.wikidata.org/wiki/Q287260","display_name":"Resource allocation","level":2,"score":0.46219998598098755},{"id":"https://openalex.org/C111873713","wikidata":"https://www.wikidata.org/wiki/Q1641413","display_name":"Job scheduler","level":3,"score":0.45320001244544983},{"id":"https://openalex.org/C55416958","wikidata":"https://www.wikidata.org/wiki/Q6206757","display_name":"Job shop scheduling","level":3,"score":0.4196000099182129},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.400299996137619},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.3418999910354614},{"id":"https://openalex.org/C206345919","wikidata":"https://www.wikidata.org/wiki/Q20380951","display_name":"Resource (disambiguation)","level":2,"score":0.32659998536109924},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.3158999979496002},{"id":"https://openalex.org/C2777958785","wikidata":"https://www.wikidata.org/wiki/Q17120940","display_name":"Resource efficiency","level":2,"score":0.3122999966144562},{"id":"https://openalex.org/C93996380","wikidata":"https://www.wikidata.org/wiki/Q44127","display_name":"Server","level":2,"score":0.3109999895095825},{"id":"https://openalex.org/C147297375","wikidata":"https://www.wikidata.org/wiki/Q6674930","display_name":"Look-ahead","level":2,"score":0.2906000018119812},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.28690001368522644},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.28519999980926514},{"id":"https://openalex.org/C2780609101","wikidata":"https://www.wikidata.org/wiki/Q17156588","display_name":"Resource management (computing)","level":2,"score":0.2847999930381775},{"id":"https://openalex.org/C2779851693","wikidata":"https://www.wikidata.org/wiki/Q183484","display_name":"Graphics processing unit","level":2,"score":0.28360000252723694},{"id":"https://openalex.org/C2984822820","wikidata":"https://www.wikidata.org/wiki/Q1123036","display_name":"Processor scheduling","level":3,"score":0.26420000195503235},{"id":"https://openalex.org/C37404715","wikidata":"https://www.wikidata.org/wiki/Q380679","display_name":"Dynamic programming","level":2,"score":0.2630999982357025},{"id":"https://openalex.org/C51332947","wikidata":"https://www.wikidata.org/wiki/Q1172305","display_name":"Shared resource","level":2,"score":0.26010000705718994},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2554999887943268}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tpds.2025.3627574","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2025.3627574","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"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 Transactions on Parallel and Distributed Systems","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3750295886","display_name":null,"funder_award_id":"LP220200893","funder_id":"https://openalex.org/F4320334704","funder_display_name":"Australian Research Council"}],"funders":[{"id":"https://openalex.org/F4320334704","display_name":"Australian Research Council","ror":"https://ror.org/05mmh0f86"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":30,"referenced_works":["https://openalex.org/W2070744914","https://openalex.org/W2107726111","https://openalex.org/W2108598243","https://openalex.org/W2145339207","https://openalex.org/W2194775991","https://openalex.org/W2622826443","https://openalex.org/W2905697694","https://openalex.org/W2912351236","https://openalex.org/W2963163009","https://openalex.org/W2963341956","https://openalex.org/W2968782695","https://openalex.org/W3017275226","https://openalex.org/W3097411828","https://openalex.org/W3100789280","https://openalex.org/W3130716829","https://openalex.org/W3165292244","https://openalex.org/W3217445637","https://openalex.org/W4285024875","https://openalex.org/W4304192541","https://openalex.org/W4323020834","https://openalex.org/W4384705353","https://openalex.org/W4385245566","https://openalex.org/W4386436407","https://openalex.org/W4386709703","https://openalex.org/W4387321503","https://openalex.org/W4394922627","https://openalex.org/W4394944658","https://openalex.org/W4399168608","https://openalex.org/W4404782703","https://openalex.org/W4406258462"],"related_works":[],"abstract_inverted_index":{"The":[0,49],"GPU":[1,41,152,165],"technology":[2],"has":[3],"significantly":[4],"aided":[5],"Deep":[6,123,131],"Learning":[7],"(DL),":[8],"especially":[9],"in":[10,105,148],"enhancing":[11],"the":[12,22,83,112,130,164,173,197,200],"performance":[13],"of":[14,52,188],"inference":[15,19,33,53,66,91],"services.":[16],"Tenants":[17],"deploy":[18],"models":[20],"on":[21],"GPU,":[23],"which":[24],"are":[25],"then":[26],"uniformly":[27],"scheduled":[28],"and":[29,58,78,117,140,170,192,202],"executed":[30],"by":[31],"an":[32,64],"serving":[34,67],"system.":[35],"In":[36,88,126,182],"resource-constrained":[37],"environments,":[38,92],"a":[39,69,107,150,186],"single":[40,151],"needs":[42],"to":[43,120,190,196],"handle":[44],"requests":[45,101],"from":[46,102],"multiple":[47,154],"tenants.":[48,155],"diver":[50],"sity":[51],"tasks,":[54],"varying":[55],"request":[56,79,118,175],"frequencies,":[57],"different":[59,168],"model":[60],"architectures":[61],"make":[62],"designing":[63],"efficient":[65],"system":[68],"significant":[70],"challenge.":[71],"Most":[72],"current":[73],"research":[74],"discusses":[75],"resource":[76,115],"allocation":[77,116,166],"batching":[80,119],"separately,":[81],"overlooking":[82],"critical":[84],"connection":[85,94,113],"between":[86,114],"them.":[87],"such":[89,106],"complex":[90],"this":[93],"is":[95],"particularly":[96],"crucial.":[97],"To":[98],"rapidly":[99,162],"process":[100],"various":[103],"tenants":[104,169],"dynamic":[108],"environment,":[109],"we":[110,128,160],"leverage":[111],"design":[121],"DRS:":[122],"Reinforcement":[124],"Scheduler.":[125],"DRS,":[127],"use":[129],"Deterministic":[132],"Policy":[133],"Gradient":[134],"(DDPG)":[135],"as":[136],"our":[137],"scheduling":[138],"algorithm":[139],"NVIDIA":[141],"Multi-Process":[142],"Service":[143],"(MPS)":[144],"for":[145,167],"spatial":[146],"parallelism":[147],"sharing":[149],"among":[153],"By":[156],"observing":[157],"environmental":[158],"information,":[159],"can":[161],"adjust":[163],"f":[171],"ind":[172],"proper":[174],"batch":[176],"size,":[177],"thereby":[178],"maintaining":[179],"high":[180],"efficiency.":[181],"experiments,":[183],"DRS":[184],"achieves":[185],"speedup":[187],"up":[189],"2.23\u00d7":[191],"24":[193],"\u00d7":[194],"compared":[195],"baselines":[198],"with":[199],"Makespan":[201],"Job":[203],"Completion":[204],"Time":[205],"(JCT)":[206],"metrics.":[207]},"counts_by_year":[],"updated_date":"2025-12-05T23:21:25.405358","created_date":"2025-10-31T00:00:00"}