{"id":"https://openalex.org/W7137861966","doi":"https://doi.org/10.1145/3802107","title":"Serving Hybrid LLM Loads with SLO Guarantees Using CPU-GPU Attention Piggybacking","display_name":"Serving Hybrid LLM Loads with SLO Guarantees Using CPU-GPU Attention Piggybacking","publication_year":2026,"publication_date":"2026-05-18","ids":{"openalex":"https://openalex.org/W7137861966","doi":"https://doi.org/10.1145/3802107"},"language":"en","primary_location":{"id":"doi:10.1145/3802107","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3802107","pdf_url":null,"source":{"id":"https://openalex.org/S4387289859","display_name":"Proceedings of the ACM on Management of Data","issn_l":"2836-6573","issn":["2836-6573"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"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 ACM on Management of Data","raw_type":"journal-article"},"type":"article","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.1145/3802107","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101324934","display_name":"Zizhao Mo","orcid":null},"institutions":[{"id":"https://openalex.org/I204512498","display_name":"University of Macau","ror":"https://ror.org/01r4q9n85","country_code":"MO","type":"education","lineage":["https://openalex.org/I204512498"]}],"countries":["MO"],"is_corresponding":false,"raw_author_name":"Zizhao Mo","raw_affiliation_strings":["University of Macau, Macau, China"],"raw_orcid":"https://orcid.org/0000-0002-3590-4400","affiliations":[{"raw_affiliation_string":"University of Macau, Macau, China","institution_ids":["https://openalex.org/I204512498"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129691543","display_name":"Junlin Chen","orcid":null},"institutions":[{"id":"https://openalex.org/I204512498","display_name":"University of Macau","ror":"https://ror.org/01r4q9n85","country_code":"MO","type":"education","lineage":["https://openalex.org/I204512498"]}],"countries":["MO"],"is_corresponding":false,"raw_author_name":"Junlin Chen","raw_affiliation_strings":["University of Macau, Macau, China"],"raw_orcid":"https://orcid.org/0009-0007-9564-1183","affiliations":[{"raw_affiliation_string":"University of Macau, Macau, China","institution_ids":["https://openalex.org/I204512498"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017636962","display_name":"Huanle Xu","orcid":"https://orcid.org/0000-0001-6657-1154"},"institutions":[{"id":"https://openalex.org/I204512498","display_name":"University of Macau","ror":"https://ror.org/01r4q9n85","country_code":"MO","type":"education","lineage":["https://openalex.org/I204512498"]}],"countries":["MO"],"is_corresponding":false,"raw_author_name":"Huanle Xu","raw_affiliation_strings":["University of Macau, Macau, China"],"raw_orcid":"https://orcid.org/0000-0001-6657-1154","affiliations":[{"raw_affiliation_string":"University of Macau, Macau, China","institution_ids":["https://openalex.org/I204512498"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129704552","display_name":"Chengzhong Xu","orcid":null},"institutions":[{"id":"https://openalex.org/I204512498","display_name":"University of Macau","ror":"https://ror.org/01r4q9n85","country_code":"MO","type":"education","lineage":["https://openalex.org/I204512498"]}],"countries":["MO"],"is_corresponding":false,"raw_author_name":"ChengZhong Xu","raw_affiliation_strings":["University of Macau, Macau, China"],"raw_orcid":"https://orcid.org/0000-0001-9480-0356","affiliations":[{"raw_affiliation_string":"University of Macau, Macau, China","institution_ids":["https://openalex.org/I204512498"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I204512498"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.27079748,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"4","issue":"3","first_page":"1","last_page":"26"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10101","display_name":"Cloud Computing and Resource Management","score":0.5339999794960022,"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":0.5339999794960022,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.10010000318288803,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.07519999891519547,"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/asynchronous-communication","display_name":"Asynchronous communication","score":0.6647999882698059},{"id":"https://openalex.org/keywords/service-provider","display_name":"Service provider","score":0.47189998626708984},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.45980000495910645},{"id":"https://openalex.org/keywords/granularity","display_name":"Granularity","score":0.41510000824928284},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.37689998745918274},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.3677999973297119},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.353300005197525},{"id":"https://openalex.org/keywords/quality-of-service","display_name":"Quality of service","score":0.3343999981880188}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8474000096321106},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.6647999882698059},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5824000239372253},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.5224999785423279},{"id":"https://openalex.org/C116537","wikidata":"https://www.wikidata.org/wiki/Q2169973","display_name":"Service provider","level":3,"score":0.47189998626708984},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.45980000495910645},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.41510000824928284},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.37689998745918274},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.3677999973297119},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.353300005197525},{"id":"https://openalex.org/C5119721","wikidata":"https://www.wikidata.org/wiki/Q220501","display_name":"Quality of service","level":2,"score":0.3343999981880188},{"id":"https://openalex.org/C51332947","wikidata":"https://www.wikidata.org/wiki/Q1172305","display_name":"Shared resource","level":2,"score":0.3276999890804291},{"id":"https://openalex.org/C2780378061","wikidata":"https://www.wikidata.org/wiki/Q25351891","display_name":"Service (business)","level":2,"score":0.313400000333786},{"id":"https://openalex.org/C93996380","wikidata":"https://www.wikidata.org/wiki/Q44127","display_name":"Server","level":2,"score":0.3059999942779541},{"id":"https://openalex.org/C49154492","wikidata":"https://www.wikidata.org/wiki/Q5300","display_name":"Central processing unit","level":2,"score":0.2897000014781952},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.2782999873161316},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.2745000123977661},{"id":"https://openalex.org/C206345919","wikidata":"https://www.wikidata.org/wiki/Q20380951","display_name":"Resource (disambiguation)","level":2,"score":0.274399995803833},{"id":"https://openalex.org/C2779089604","wikidata":"https://www.wikidata.org/wiki/Q7169333","display_name":"Permission","level":2,"score":0.2727000117301941},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2671999931335449},{"id":"https://openalex.org/C2778160497","wikidata":"https://www.wikidata.org/wiki/Q869830","display_name":"Service-level agreement","level":3,"score":0.2662000060081482},{"id":"https://openalex.org/C2777480716","wikidata":"https://www.wikidata.org/wiki/Q23582796","display_name":"Resource consumption","level":2,"score":0.26249998807907104}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3802107","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3802107","pdf_url":null,"source":{"id":"https://openalex.org/S4387289859","display_name":"Proceedings of the ACM on Management of Data","issn_l":"2836-6573","issn":["2836-6573"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"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 ACM on Management of Data","raw_type":"journal-article"},{"id":"pmh:oai:arXiv.org:2603.12831","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2603.12831","pdf_url":"https://arxiv.org/pdf/2603.12831","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"doi:10.1145/3802107","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3802107","pdf_url":null,"source":{"id":"https://openalex.org/S4387289859","display_name":"Proceedings of the ACM on Management of Data","issn_l":"2836-6573","issn":["2836-6573"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"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 ACM on Management of Data","raw_type":"journal-article"},"sustainable_development_goals":[{"score":0.43205708265304565,"id":"https://metadata.un.org/sdg/8","display_name":"Decent work and economic growth"}],"awards":[],"funders":[{"id":"https://openalex.org/F4320311133","display_name":"United Mitochondrial Disease Foundation","ror":"https://ror.org/0528q0t18"},{"id":"https://openalex.org/F4320322841","display_name":"Universidade de Macau","ror":"https://ror.org/01r4q9n85"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Nowadays,":[0],"service":[1,15,43,81],"providers":[2],"often":[3],"deploy":[4],"multiple":[5],"types":[6],"of":[7,45,78,88,131],"LLM":[8,100],"services":[9,28,48,133,192],"within":[10],"shared":[11],"clusters.":[12],"While":[13],"the":[14,42,75,79,85,89,121,128,137,186],"co-location":[16],"improves":[17,185],"resource":[18,67],"utilization,":[19],"it":[20],"introduces":[21],"significant":[22],"interference":[23,113],"risks":[24],"for":[25,35,190],"latency-sensitive":[26,80],"(LS)":[27],"--":[29,38],"which":[30,125],"have":[31],"strict":[32],"SLO":[33,76,187],"requirements":[34],"inference":[36],"latency":[37],"and":[39,82,108,114,147],"severely":[40],"constrains":[41],"capacity":[44],"best-effort":[46,90],"(BE)":[47],"due":[49],"to":[50,64,111,118,134,167,169,195,204,207],"limited":[51],"available":[52],"memory.":[53],"To":[54],"address":[55],"interference,":[56],"existing":[57],"systems":[58],"typically":[59],"rely":[60],"on":[61,136],"reserving":[62],"headroom":[63],"constrain":[65],"BE":[66,132,199],"usage.":[68],"However,":[69],"this":[70,93],"approach's":[71],"coarse":[72],"granularity":[73],"compromises":[74],"compliance":[77],"unnecessarily":[83],"restricts":[84],"generation":[86],"potential":[87],"service.":[91],"In":[92],"paper,":[94],"we":[95],"propose":[96],"OmniServe,":[97],"a":[98,162],"novel":[99],"serving":[101,200],"system":[102],"that":[103,183],"efficiently":[104],"harnesses":[105],"both":[106],"CPU":[107,146],"GPU":[109,148],"resources":[110],"mitigate":[112],"improve":[115],"throughput.":[116],"Central":[117],"OmniServe":[119,160,184],"is":[120],"Attention":[122,129,157],"Piggybacking":[123],"mechanism,":[124],"effectively":[126],"offloads":[127],"computation":[130,176],"CPUs":[135],"fly.":[138],"This":[139],"mechanism":[140],"also":[141],"facilitates":[142],"asynchronous":[143],"communication":[144],"between":[145],"streams,":[149],"preventing":[150],"GPUs":[151],"from":[152],"being":[153],"blocked":[154],"while":[155,197],"aggregating":[156],"results.":[158],"Additionally,":[159],"incorporates":[161],"dynamic":[163],"batching":[164],"control":[165],"policy":[166],"adapt":[168],"fluctuating":[170],"request":[171],"arrivals,":[172],"facilitating":[173],"Dense":[174],"module":[175],"using":[177],"layer-wise":[178],"batching.":[179],"Experimental":[180],"results":[181],"show":[182],"attainment":[188],"rate":[189],"LS":[191],"by":[193,202],"up":[194,203],"1.48\u00d7":[196],"enhancing":[198],"throughput":[201],"9.85\u00d7":[205],"compared":[206],"state-of-the-art":[208],"systems.":[209]},"counts_by_year":[],"updated_date":"2026-08-26T07:47:46.906454","created_date":"2026-03-18T00:00:00"}
