{"id":"https://openalex.org/W4416352716","doi":"https://doi.org/10.1145/3745756.3809234","title":"KVSwap: Disk-aware KV Cache Offloading for Long-Context On-device Inference","display_name":"KVSwap: Disk-aware KV Cache Offloading for Long-Context On-device Inference","publication_year":2026,"publication_date":"2026-05-29","ids":{"openalex":"https://openalex.org/W4416352716","doi":"https://doi.org/10.1145/3745756.3809234"},"language":"en","primary_location":{"id":"doi:10.1145/3745756.3809234","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3745756.3809234","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 24th Annual International Conference on Mobile Systems, Applications and Services","raw_type":"proceedings-article"},"type":"article","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3745756.3809234","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100320805","display_name":"Huawei Zhang","orcid":"https://orcid.org/0000-0003-1021-3184"},"institutions":[{"id":"https://openalex.org/I130828816","display_name":"University of Leeds","ror":"https://ror.org/024mrxd33","country_code":"GB","type":"education","lineage":["https://openalex.org/I130828816"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Huawei Zhang","raw_affiliation_strings":["University of Leeds, Leeds, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0003-1021-3184","affiliations":[{"raw_affiliation_string":"University of Leeds, Leeds, United Kingdom","institution_ids":["https://openalex.org/I130828816"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5095381671","display_name":"Chunwei Xia","orcid":"https://orcid.org/0000-0003-2014-5453"},"institutions":[{"id":"https://openalex.org/I130828816","display_name":"University of Leeds","ror":"https://ror.org/024mrxd33","country_code":"GB","type":"education","lineage":["https://openalex.org/I130828816"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Chunwei Xia","raw_affiliation_strings":["University of Leeds, Leeds, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0003-2014-5453","affiliations":[{"raw_affiliation_string":"University of Leeds, Leeds, United Kingdom","institution_ids":["https://openalex.org/I130828816"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5082429626","display_name":"Zheng Wang","orcid":"https://orcid.org/0000-0001-8305-0676"},"institutions":[{"id":"https://openalex.org/I130828816","display_name":"University of Leeds","ror":"https://ror.org/024mrxd33","country_code":"GB","type":"education","lineage":["https://openalex.org/I130828816"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Zheng Wang","raw_affiliation_strings":["University of Leeds, Leeds, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0001-6157-0662","affiliations":[{"raw_affiliation_string":"University of Leeds, Leeds, United Kingdom","institution_ids":["https://openalex.org/I130828816"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.00242669,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"723","last_page":"736"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T14347","display_name":"Big Data and Digital Economy","score":0.22300000488758087,"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/T14347","display_name":"Big Data and Digital Economy","score":0.22300000488758087,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.14720000326633453,"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/T10273","display_name":"IoT and Edge/Fog Computing","score":0.09889999777078629,"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/cache","display_name":"Cache","score":0.794700026512146},{"id":"https://openalex.org/keywords/mobile-device","display_name":"Mobile device","score":0.47620001435279846},{"id":"https://openalex.org/keywords/cpu-cache","display_name":"CPU cache","score":0.4221000075340271},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.42179998755455017},{"id":"https://openalex.org/keywords/cache-algorithms","display_name":"Cache algorithms","score":0.4189000129699707},{"id":"https://openalex.org/keywords/metadata","display_name":"Metadata","score":0.3991999924182892},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.37720000743865967},{"id":"https://openalex.org/keywords/cache-pollution","display_name":"Cache pollution","score":0.37610000371932983},{"id":"https://openalex.org/keywords/page-cache","display_name":"Page cache","score":0.37380000948905945}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8776999711990356},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.794700026512146},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.4796999990940094},{"id":"https://openalex.org/C186967261","wikidata":"https://www.wikidata.org/wiki/Q5082128","display_name":"Mobile device","level":2,"score":0.47620001435279846},{"id":"https://openalex.org/C189783530","wikidata":"https://www.wikidata.org/wiki/Q352090","display_name":"CPU cache","level":3,"score":0.4221000075340271},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.42179998755455017},{"id":"https://openalex.org/C38556500","wikidata":"https://www.wikidata.org/wiki/Q13404475","display_name":"Cache algorithms","level":4,"score":0.4189000129699707},{"id":"https://openalex.org/C93518851","wikidata":"https://www.wikidata.org/wiki/Q180160","display_name":"Metadata","level":2,"score":0.3991999924182892},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.39739999175071716},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.37720000743865967},{"id":"https://openalex.org/C113166858","wikidata":"https://www.wikidata.org/wiki/Q5015981","display_name":"Cache pollution","level":5,"score":0.37610000371932983},{"id":"https://openalex.org/C36340418","wikidata":"https://www.wikidata.org/wiki/Q7124288","display_name":"Page cache","level":5,"score":0.37380000948905945},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3684999942779541},{"id":"https://openalex.org/C201148951","wikidata":"https://www.wikidata.org/wiki/Q5015976","display_name":"Cache coloring","level":4,"score":0.3626999855041504},{"id":"https://openalex.org/C167713795","wikidata":"https://www.wikidata.org/wiki/Q352090","display_name":"Smart Cache","level":5,"score":0.33899998664855957},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3379000127315521},{"id":"https://openalex.org/C49154492","wikidata":"https://www.wikidata.org/wiki/Q5300","display_name":"Central processing unit","level":2,"score":0.31310001015663147},{"id":"https://openalex.org/C202623185","wikidata":"https://www.wikidata.org/wiki/Q375176","display_name":"Disk buffer","level":3,"score":0.3046000003814697},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.2985000014305115},{"id":"https://openalex.org/C144543869","wikidata":"https://www.wikidata.org/wiki/Q2738570","display_name":"Mobile computing","level":2,"score":0.2791999876499176},{"id":"https://openalex.org/C170858558","wikidata":"https://www.wikidata.org/wiki/Q1394144","display_name":"Automatic summarization","level":2,"score":0.27160000801086426},{"id":"https://openalex.org/C59687516","wikidata":"https://www.wikidata.org/wiki/Q5015938","display_name":"Cache-oblivious algorithm","level":5,"score":0.267300009727478},{"id":"https://openalex.org/C82687282","wikidata":"https://www.wikidata.org/wiki/Q66221","display_name":"Auxiliary memory","level":2,"score":0.2635999917984009},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.2623000144958496},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.2619999945163727},{"id":"https://openalex.org/C53833338","wikidata":"https://www.wikidata.org/wiki/Q1061424","display_name":"Context switch","level":2,"score":0.25760000944137573},{"id":"https://openalex.org/C93996380","wikidata":"https://www.wikidata.org/wiki/Q44127","display_name":"Server","level":2,"score":0.2572000026702881},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.25380000472068787}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.1145/3745756.3809234","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3745756.3809234","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 24th Annual International Conference on Mobile Systems, Applications and Services","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2511.11907","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2511.11907","pdf_url":"https://arxiv.org/pdf/2511.11907","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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:eprints.whiterose.ac.uk:240121","is_oa":true,"landing_page_url":"https://orcid.org/0000-0001-6157-0662>","pdf_url":"https://eprints.whiterose.ac.uk/id/eprint/240121/1/kvswap.pdf","source":{"id":"https://openalex.org/S4306400854","display_name":"White Rose Research Online (University of Leeds, The University of Sheffield, University of York)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I2800616092","host_organization_name":"White Rose University Consortium","host_organization_lineage":["https://openalex.org/I2800616092"],"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":"NonPeerReviewed"},{"id":"doi:10.48550/arxiv.2511.11907","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2511.11907","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.1145/3745756.3809234","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3745756.3809234","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 24th Annual International Conference on Mobile Systems, Applications and Services","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Language":[0],"models":[1],"(LMs)":[2],"underpin":[3],"emerging":[4],"mobile":[5,84],"and":[6,12,15,35,57,83,89,98,152,168],"embedded":[7,82],"AI":[8],"applications":[9],"like":[10],"meeting":[11],"video":[13],"summarization":[14],"document":[16],"analysis,":[17],"which":[18,142],"often":[19],"require":[20],"processing":[21],"multiple":[22],"long-context":[23,39],"inputs.":[24],"Running":[25],"an":[26],"LM":[27],"locally":[28],"on-device":[29],"improves":[30],"privacy,":[31],"enables":[32],"offline":[33],"use,":[34],"reduces":[36],"cost,":[37],"but":[38],"inference":[40],"quickly":[41],"hits":[42],"a":[43,95,111],"\\emph{memory":[44],"capacity":[45],"wall}":[46],"as":[47],"the":[48,87,99,130],"key-value":[49],"(KV)":[50],"cache":[51,68,132,186],"grows":[52],"linearly":[53],"with":[54,148],"context":[55],"length":[56],"batch":[58],"size.":[59],"Existing":[60],"KV-cache":[61],"offloading":[62,187],"schemes":[63],"are":[64,78],"designed":[65],"to":[66,73,140,144,156],"transfer":[67],"data":[69],"from":[70],"GPU":[71,90],"memory":[72,97,121,177],"CPU":[74,88],"memory;":[75],"however,":[76],"they":[77],"not":[79],"suitable":[80],"for":[81,115],"systems,":[85],"where":[86],"(or":[91],"NPU)":[92],"typically":[93],"share":[94],"unified":[96],"non-volatile":[100],"secondary":[101],"storage":[102,158,169],"(disk)":[103],"offers":[104],"limited":[105],"I/O":[106],"bandwidth.":[107],"We":[108],"present":[109],"KVSwap,":[110],"software":[112],"framework":[113],"tailored":[114],"local":[116],"devices":[117],"that":[118,164],"achieves":[119],"high":[120],"efficiency":[122],"while":[123,179],"effectively":[124],"leveraging":[125],"disk":[126,150],"storage.":[127],"KVSwap":[128,171],"stores":[129],"full":[131],"on":[133],"disk,":[134],"uses":[135],"highly":[136],"compact":[137],"in-memory":[138],"metadata":[139],"predict":[141],"entries":[143],"preload,":[145],"overlaps":[146],"computation":[147],"hardware-aware":[149],"access,":[151],"orchestrates":[153],"read":[154],"patterns":[155],"match":[157],"device":[159],"characteristics.":[160],"Our":[161],"evaluation":[162],"shows":[163],"across":[165],"representative":[166],"LMs":[167],"types,":[170],"delivers":[172],"higher":[173],"throughput":[174],"under":[175],"tight":[176],"budgets":[178],"maintaining":[180],"generation":[181],"quality":[182],"over":[183],"existing":[184],"KV":[185],"schemes.":[188]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-11-19T00:00:00"}
