{"id":"https://openalex.org/W4409060278","doi":"https://doi.org/10.1145/3721146.3721956","title":"Priority-Aware Preemptive Scheduling for Mixed-Priority Workloads in MoE Inference","display_name":"Priority-Aware Preemptive Scheduling for Mixed-Priority Workloads in MoE Inference","publication_year":2025,"publication_date":"2025-03-30","ids":{"openalex":"https://openalex.org/W4409060278","doi":"https://doi.org/10.1145/3721146.3721956"},"language":"en","primary_location":{"id":"doi:10.1145/3721146.3721956","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721146.3721956","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 5th Workshop on Machine Learning and Systems","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3721146.3721956","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5001439830","display_name":"Mohammad Siavashi","orcid":"https://orcid.org/0000-0002-2600-9025"},"institutions":[{"id":"https://openalex.org/I86987016","display_name":"KTH Royal Institute of Technology","ror":"https://ror.org/026vcq606","country_code":"SE","type":"education","lineage":["https://openalex.org/I86987016"]}],"countries":["SE"],"is_corresponding":false,"raw_author_name":"Mohammad Siavashi","raw_affiliation_strings":["KTH Royal Institute of Technology, Stockholm, Sweden"],"raw_orcid":"https://orcid.org/0000-0002-2600-9025","affiliations":[{"raw_affiliation_string":"KTH Royal Institute of Technology, Stockholm, Sweden","institution_ids":["https://openalex.org/I86987016"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5116870962","display_name":"Faezeh Keshmiri Dindarloo","orcid":"https://orcid.org/0009-0008-2161-0437"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Faezeh Keshmiri Dindarloo","raw_affiliation_strings":["Unaffiliated, Stockholm, Sweden"],"raw_orcid":"https://orcid.org/0009-0008-2161-0437","affiliations":[{"raw_affiliation_string":"Unaffiliated, Stockholm, Sweden","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048036382","display_name":"Dejan Kosti\u0107","orcid":"https://orcid.org/0000-0002-1256-1070"},"institutions":[{"id":"https://openalex.org/I86987016","display_name":"KTH Royal Institute of Technology","ror":"https://ror.org/026vcq606","country_code":"SE","type":"education","lineage":["https://openalex.org/I86987016"]}],"countries":["SE"],"is_corresponding":false,"raw_author_name":"Dejan Kostic","raw_affiliation_strings":["KTH Royal Institute of Technology, Stockholm, Sweden"],"raw_orcid":"https://orcid.org/0000-0002-1256-1070","affiliations":[{"raw_affiliation_string":"KTH Royal Institute of Technology, Stockholm, Sweden","institution_ids":["https://openalex.org/I86987016"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5058517041","display_name":"Marco Chiesa","orcid":"https://orcid.org/0000-0002-9675-9729"},"institutions":[{"id":"https://openalex.org/I86987016","display_name":"KTH Royal Institute of Technology","ror":"https://ror.org/026vcq606","country_code":"SE","type":"education","lineage":["https://openalex.org/I86987016"]}],"countries":["SE"],"is_corresponding":false,"raw_author_name":"Marco Chiesa","raw_affiliation_strings":["KTH Royal Institute of Technology, Stockholm, Sweden"],"raw_orcid":"https://orcid.org/0000-0002-9675-9729","affiliations":[{"raw_affiliation_string":"KTH Royal Institute of Technology, Stockholm, Sweden","institution_ids":["https://openalex.org/I86987016"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.8048,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.91264128,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"132","last_page":"138"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10679","display_name":"Service-Oriented Architecture and Web Services","score":0.995199978351593,"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/T10679","display_name":"Service-Oriented Architecture and Web Services","score":0.995199978351593,"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":0.9890999794006348,"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/T10927","display_name":"Access Control and Trust","score":0.9869999885559082,"subfield":{"id":"https://openalex.org/subfields/3312","display_name":"Sociology and Political Science"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8169604539871216},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.6002494692802429},{"id":"https://openalex.org/keywords/deadline-monotonic-scheduling","display_name":"Deadline-monotonic scheduling","score":0.5584885478019714},{"id":"https://openalex.org/keywords/preemption","display_name":"Preemption","score":0.5152931213378906},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4883532226085663},{"id":"https://openalex.org/keywords/earliest-deadline-first-scheduling","display_name":"Earliest deadline first scheduling","score":0.4847792983055115},{"id":"https://openalex.org/keywords/fixed-priority-pre-emptive-scheduling","display_name":"Fixed-priority pre-emptive scheduling","score":0.47406524419784546},{"id":"https://openalex.org/keywords/rate-monotonic-scheduling","display_name":"Rate-monotonic scheduling","score":0.35419440269470215},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.31373119354248047},{"id":"https://openalex.org/keywords/dynamic-priority-scheduling","display_name":"Dynamic priority scheduling","score":0.3076821565628052},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.27583304047584534},{"id":"https://openalex.org/keywords/quality-of-service","display_name":"Quality of service","score":0.11997425556182861},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.11846891045570374},{"id":"https://openalex.org/keywords/operations-management","display_name":"Operations management","score":0.06324365735054016},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.0578344464302063}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8169604539871216},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6002494692802429},{"id":"https://openalex.org/C76774629","wikidata":"https://www.wikidata.org/wiki/Q1146103","display_name":"Deadline-monotonic scheduling","level":5,"score":0.5584885478019714},{"id":"https://openalex.org/C206952183","wikidata":"https://www.wikidata.org/wiki/Q1193100","display_name":"Preemption","level":2,"score":0.5152931213378906},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4883532226085663},{"id":"https://openalex.org/C32310161","wikidata":"https://www.wikidata.org/wiki/Q666365","display_name":"Earliest deadline first scheduling","level":5,"score":0.4847792983055115},{"id":"https://openalex.org/C122141398","wikidata":"https://www.wikidata.org/wiki/Q5456330","display_name":"Fixed-priority pre-emptive scheduling","level":5,"score":0.47406524419784546},{"id":"https://openalex.org/C127456818","wikidata":"https://www.wikidata.org/wiki/Q238879","display_name":"Rate-monotonic scheduling","level":4,"score":0.35419440269470215},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.31373119354248047},{"id":"https://openalex.org/C107568181","wikidata":"https://www.wikidata.org/wiki/Q5319000","display_name":"Dynamic priority scheduling","level":3,"score":0.3076821565628052},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.27583304047584534},{"id":"https://openalex.org/C5119721","wikidata":"https://www.wikidata.org/wiki/Q220501","display_name":"Quality of service","level":2,"score":0.11997425556182861},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.11846891045570374},{"id":"https://openalex.org/C21547014","wikidata":"https://www.wikidata.org/wiki/Q1423657","display_name":"Operations management","level":1,"score":0.06324365735054016},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0578344464302063}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1145/3721146.3721956","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721146.3721956","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 5th Workshop on Machine Learning and Systems","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2503.09304","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.09304","pdf_url":"https://arxiv.org/pdf/2503.09304","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"pmh:oai:DiVA.org:kth-364688","is_oa":true,"landing_page_url":"http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-364688","pdf_url":null,"source":{"id":"https://openalex.org/S4306400653","display_name":"Diva portal (Dalarna University Library)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":{"id":"doi:10.1145/3721146.3721956","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721146.3721956","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 5th Workshop on Machine Learning and Systems","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G6585254729","display_name":"SEMLA: Securing Enterprises via Machine-Learning-based Automation","funder_award_id":"2023-03003","funder_id":"https://openalex.org/F4320321030","funder_display_name":"VINNOVA"},{"id":"https://openalex.org/G6894485844","display_name":"ResoNet: Resilient Optimized Network Synthesis","funder_award_id":"2021-04212","funder_id":"https://openalex.org/F4320322581","funder_display_name":"Vetenskapsr\u00e5det"}],"funders":[{"id":"https://openalex.org/F4320321030","display_name":"VINNOVA","ror":"https://ror.org/01kd5m353"},{"id":"https://openalex.org/F4320322581","display_name":"Vetenskapsr\u00e5det","ror":"https://ror.org/03zttf063"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":9,"referenced_works":["https://openalex.org/W95608104","https://openalex.org/W2761028280","https://openalex.org/W2899575547","https://openalex.org/W2963929190","https://openalex.org/W3155584966","https://openalex.org/W4401176373","https://openalex.org/W6600300155","https://openalex.org/W6604186633","https://openalex.org/W6655411870"],"related_works":["https://openalex.org/W2377321505","https://openalex.org/W2962323017","https://openalex.org/W2393789493","https://openalex.org/W4300092030","https://openalex.org/W2169142010","https://openalex.org/W1827860574","https://openalex.org/W2386407620","https://openalex.org/W2165972904","https://openalex.org/W2167574351","https://openalex.org/W2374555443"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"have":[3],"revolutionized":[4],"natural":[5],"language":[6],"processing,":[7],"yet":[8],"serving":[9],"them":[10],"efficiently":[11],"in":[12],"data":[13],"centers":[14],"remains":[15],"challenging":[16],"due":[17],"to":[18,86,122,129,144],"mixed":[19],"workloads":[20],"comprising":[21],"latency-sensitive":[22],"(LS)":[23],"and":[24,116,153],"best-effort":[25],"(BE)":[26],"jobs.":[27,43],"Existing":[28],"inference":[29,49],"systems":[30],"employ":[31],"iteration-level":[32,80],"first-come-first-served":[33],"scheduling,":[34],"causing":[35],"head-of-line":[36],"blocking":[37],"when":[38],"BE":[39,69],"jobs":[40,88],"delay":[41],"LS":[42,74,109,139],"We":[44],"introduce":[45],"QLLM,":[46],"a":[47,59],"novel":[48],"system":[50],"designed":[51],"for":[52],"Mixture":[53],"of":[54,114],"Experts":[55],"(MoE)":[56],"models,":[57],"featuring":[58],"fine-grained,":[60],"priority-aware":[61],"preemptive":[62],"scheduler.":[63],"QLLM":[64,103,149],"enables":[65],"expert-level":[66],"preemption,":[67],"deferring":[68],"job":[70],"execution":[71],"while":[72],"minimizing":[73],"time-to-first-token":[75],"(TTFT).":[76],"Our":[77],"approach":[78],"removes":[79],"scheduling":[81],"constraints,":[82],"enabling":[83],"the":[84,118,126,133],"scheduler":[85],"preempt":[87],"at":[89,120],"any":[90],"layer":[91],"based":[92],"on":[93,96],"priority.":[94],"Evaluations":[95],"an":[97,112],"Nvidia":[98],"A100":[99],"GPU":[100],"show":[101],"that":[102],"significantly":[104],"improves":[105],"performance.":[106],"It":[107],"reduces":[108],"TTFT":[110],"by":[111,142],"average":[113],"65.5\u00d7":[115],"meets":[117],"SLO":[119],"up":[121,143],"7":[123],"requests/sec,":[124],"whereas":[125],"baseline":[127],"fails":[128],"do":[130],"so":[131],"under":[132],"tested":[134],"workload.":[135],"Additionally,":[136],"it":[137],"cuts":[138],"turnaround":[140],"time":[141],"12.8\u00d7":[145],"without":[146],"impacting":[147],"throughput.":[148],"is":[150],"modular,":[151],"extensible,":[152],"seamlessly":[154],"integrates":[155],"with":[156],"Hugging":[157],"Face":[158],"MoE":[159],"models.":[160]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-08-01T09:00:35.917206","created_date":"2025-10-10T00:00:00"}