{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T06:13:33Z","timestamp":1773814413763,"version":"3.50.1"},"reference-count":44,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"4","license":[{"start":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T00:00:00Z","timestamp":1775001600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T00:00:00Z","timestamp":1775001600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T00:00:00Z","timestamp":1775001600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Comput."],"published-print":{"date-parts":[[2026,4]]},"DOI":"10.1109\/tc.2026.3656215","type":"journal-article","created":{"date-parts":[[2026,1,28]],"date-time":"2026-01-28T20:58:51Z","timestamp":1769633931000},"page":"1309-1322","source":"Crossref","is-referenced-by-count":0,"title":["Bauhaus: Restructuring Vector Database for LLM Retrieval on CXL-Based Tiered Memory"],"prefix":"10.1109","volume":"75","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-0809-317X","authenticated-orcid":false,"given":"Kyungbin","family":"Kim","sequence":"first","affiliation":[{"name":"School of Electrical Engineering, Korea University, Seoul, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sungsu","family":"Ahn","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea University, Seoul, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Wonjung","family":"Jeong","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea University, Seoul, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jongmin","family":"Kim","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea University, Seoul, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-5803-361X","authenticated-orcid":false,"given":"Sangun","family":"Choi","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea University, Seoul, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Minseong","family":"Gil","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea University, Seoul, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Minseong","family":"Kim","sequence":"additional","affiliation":[{"name":"SK hynix Inc., Icheon, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Dongha","family":"Jung","sequence":"additional","affiliation":[{"name":"SK hynix Inc., Icheon, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yunjay","family":"Hong","sequence":"additional","affiliation":[{"name":"SK hynix Inc., Icheon, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Haekang","family":"Jung","sequence":"additional","affiliation":[{"name":"SK hynix Inc., Icheon, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6442-3705","authenticated-orcid":false,"given":"Yunho","family":"Oh","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea University, Seoul, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","first-page":"9459","article-title":"Retrieval-augmented generation for knowledge-intensive NLP tasks","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"33","author":"Lewis","year":"2020"},{"key":"ref2","first-page":"2206","article-title":"Improving language models by retrieving from trillions of tokens","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Borgeaud","year":"2022"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i24.34743"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00530"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1145\/3669900"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/LCA.2022.3226482"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA59077.2024.00036"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/FCCM57271.2023.00017"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2023.3250407"},{"key":"ref10","first-page":"105","article-title":"Demystifying CXL memory with genuine CXL-ready systems and devices","volume-title":"Proc. 56th Annu. IEEE\/ACM Int. Symp. Microarchit.","author":"Sun","year":"2023"},{"key":"ref11","article-title":"The Hitchhiker\u2019s guide to programming and optimizing CXL-based heterogeneous systems","author":"Wang","year":"2024"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS64566.2025.00097"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2018.2889473"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3695053.3731032"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3582016.3582063"},{"key":"ref16","article-title":"Damon: Data access monitor","author":"Park","year":"2020"},{"key":"ref17","article-title":"AutoNUMA: The other approach to NUMA scheduling","author":"Corbet","year":"2012"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1145\/3600006.3613167"},{"key":"ref19","first-page":"817","article-title":"{FlexMem}: Adaptive page profiling and migration for tiered memory","volume-title":"Proc. USENIX Annu. Tech. Conf. (USENIX ATC)","author":"Xu","year":"2024"},{"key":"ref20","volume-title":"Intel\u00ae 64 and IA-32 Architectures Software Developer\u2019s Manual","volume":"3B","year":"2016"},{"key":"ref21","volume-title":"Understanding the Linux Virtual Memory Manager","volume":"352","author":"Gorman","year":"2004"},{"key":"ref22","article-title":"HMSDK: Heterogeneous memory software development kit.\u201d","year":"2026"},{"key":"ref23","article-title":"SMDK: Scalable memory development kit.\u201d","author":"Samsung","year":"2025"},{"key":"ref24","article-title":"Retrieval-augmented generation for large language models: A survey","author":"Gao","year":"2023"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-63646-2_29"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1810.04805"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/tbdata.2025.3618474"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.14778\/3303753.3303754"},{"key":"ref29","article-title":"Qdrant: High-performance, massive-scale vector database and vector search engine for the next generation of AI.\u201d","year":"2025"},{"issue":"140","key":"ref30","first-page":"1","article-title":"Exploring the limits of transfer learning with a unified text-to-text transformer","volume":"21","author":"Raffel","year":"2020","journal-title":"J. Mach. Learn. Res."},{"key":"ref31","first-page":"5776","article-title":"MiniLM: Deep self-attention distillation for task-agnostic compression of pre-trained transformers","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"33","author":"Wang","year":"2020"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0493"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00276"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/3768628"},{"key":"ref35","article-title":"An adaptive vector index partitioning scheme for low-latency RAG pipeline","author":"Kim","year":"2025"},{"key":"ref36","first-page":"19","article-title":"Nomad:{Non-Exclusive} memory tiering via transactional page migration","volume-title":"Proc. 18th USENIX Symp. Oper. Syst. Des. Implementation (OSDI)","author":"Xiang","year":"2024"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1145\/3626772.3657878"},{"key":"ref38","article-title":"MEDLINE\/PubMed Baseline Repository (MBR)","year":"2023"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-023-02068-4"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1109\/ISPASS.2014.6844459"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1145\/3689031.3717462"},{"key":"ref42","first-page":"37","article-title":"Managing memory tiers with {CXL} in virtualized environments","volume-title":"Proc. 18th USENIX Symp. Oper. Syst. Des. Implementation (OSDI)","author":"Zhong","year":"2024"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/TC.2020.2973134"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2789"}],"container-title":["IEEE Transactions on Computers"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/12\/11435995\/11363378.pdf?arnumber=11363378","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T05:18:52Z","timestamp":1773811132000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11363378\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4]]},"references-count":44,"journal-issue":{"issue":"4"},"URL":"https:\/\/doi.org\/10.1109\/tc.2026.3656215","relation":{},"ISSN":["0018-9340","1557-9956","2326-3814"],"issn-type":[{"value":"0018-9340","type":"print"},{"value":"1557-9956","type":"electronic"},{"value":"2326-3814","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4]]}}}