{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,6]],"date-time":"2026-06-06T00:31:29Z","timestamp":1780705889899,"version":"3.54.1"},"reference-count":63,"publisher":"Association for Computing Machinery (ACM)","issue":"3","license":[{"start":{"date-parts":[[2021,8,16]],"date-time":"2021-08-16T00:00:00Z","timestamp":1629072000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"name":"National Key Research and Development Program of China","award":["2020YFB2104100"],"award-info":[{"award-number":["2020YFB2104100"]}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["61732010 and 61925206"],"award-info":[{"award-number":["61732010 and 61925206"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100003816","name":"Huawei Technologies","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100003816","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Storage"],"published-print":{"date-parts":[[2021,8,31]]},"abstract":"<jats:p>\n            <jats:bold>RDMA<\/jats:bold>\n            (\n            <jats:bold>Remote Direct Memory Access<\/jats:bold>\n            ) has gained considerable interests in network-attached in-memory key-value stores. However, traversing the remote tree-based index in ordered key-value stores with RDMA becomes a critical obstacle, causing an order-of-magnitude slowdown and limited scalability due to multiple round trips. Using index cache with conventional wisdom\u2014caching partial data and traversing them locally\u2014usually leads to limited effect because of unavoidable capacity misses, massive random accesses, and costly cache invalidations.\n          <\/jats:p>\n          <jats:p>\n            We argue that the\n            <jats:bold>machine learning<\/jats:bold>\n            (ML) model is a perfect cache structure for the tree-based index, termed\n            <jats:italic>learned cache<\/jats:italic>\n            . Based on it, we design and implement\n            <jats:sc>XStore<\/jats:sc>\n            , an RDMA-based ordered key-value store with a new hybrid architecture that retains a tree-based index at the server to perform dynamic workloads (e.g., inserts) and leverages a learned cache at the client to perform static workloads (e.g., gets and scans). The key idea is to decouple ML model retraining from index updating by maintaining a layer of indirection from logical to actual positions of key-value pairs. It allows a stale learned cache to continue predicting a correct position for a lookup key.\n            <jats:sc>XStore<\/jats:sc>\n            ensures correctness using a validation mechanism with a fallback path and further uses speculative execution to minimize the cost of cache misses. Evaluations with YCSB benchmarks and production workloads show that a single\n            <jats:sc>XStore<\/jats:sc>\n            server can achieve over 80 million read-only requests per second. This number outperforms state-of-the-art RDMA-based ordered key-value stores (namely, DrTM-Tree, Cell, and eRPC+Masstree) by up to 5.9\u00d7 (from 3.7\u00d7). For workloads with inserts,\n            <jats:sc>XStore<\/jats:sc>\n            still provides up to 3.5\u00d7 (from 2.7\u00d7) throughput speedup, achieving 53M reqs\/s. The learned cache can also reduce client-side memory usage and further provides an efficient memory-performance tradeoff, e.g., saving 99% memory at the cost of 20% peak throughput.\n          <\/jats:p>","DOI":"10.1145\/3468520","type":"journal-article","created":{"date-parts":[[2021,8,17]],"date-time":"2021-08-17T00:10:00Z","timestamp":1629159000000},"page":"1-32","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":35,"title":["<scp>XStore<\/scp>\n            : Fast RDMA-Based Ordered Key-Value Store Using Remote Learned Cache"],"prefix":"10.1145","volume":"17","author":[{"given":"Xingda","family":"Wei","sequence":"first","affiliation":[{"name":"Shanghai Jiao Tong University, China and Shanghai Artificial Intelligence Laboratory, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rong","family":"Chen","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University, China and Shanghai Artificial Intelligence Laboratory, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Haibo","family":"Chen","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University, China and Domain-Specific Operating Systems, Ministry of Education, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Binyu","family":"Zang","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University, China and Domain-Specific Operating Systems, Ministry of Education, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2021,8,16]]},"reference":[{"key":"e_1_2_1_1_1","unstructured":"2021. Intel\u2019s Math Kernel Library. (2021). https:\/\/software.intel.com\/content\/www\/us\/en\/develop\/tools\/math-kernel-library.html.  2021. Intel\u2019s Math Kernel Library. (2021). https:\/\/software.intel.com\/content\/www\/us\/en\/develop\/tools\/math-kernel-library.html."},{"key":"e_1_2_1_2_1","volume-title":"https:\/\/memcached.org\/","year":"2021"},{"key":"e_1_2_1_3_1","volume-title":"https:\/\/aws.amazon.com\/public-datasets\/osm","author":"StreetMap","year":"2021"},{"key":"e_1_2_1_4_1","doi-asserted-by":"publisher","DOI":"10.5555\/3026877.3026899"},{"key":"e_1_2_1_5_1","doi-asserted-by":"publisher","DOI":"10.1145\/3317550.3321433"},{"key":"e_1_2_1_6_1","doi-asserted-by":"publisher","DOI":"10.1145\/2318857.2254766"},{"key":"e_1_2_1_7_1","doi-asserted-by":"publisher","DOI":"10.1109\/L-CA.2006.18"},{"key":"e_1_2_1_8_1","doi-asserted-by":"publisher","DOI":"10.5555\/2535461.2535468"},{"key":"e_1_2_1_9_1","doi-asserted-by":"publisher","DOI":"10.14778\/3229863.3229872"},{"key":"e_1_2_1_10_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2017.2729545"},{"key":"e_1_2_1_11_1","doi-asserted-by":"publisher","DOI":"10.1145\/3092701"},{"key":"e_1_2_1_12_1","doi-asserted-by":"publisher","DOI":"10.1145\/3302424.3303968"},{"key":"e_1_2_1_13_1","doi-asserted-by":"publisher","DOI":"10.1145\/2901318.2901349"},{"key":"e_1_2_1_14_1","unstructured":"Brian F. Cooper. 2021. YCSB Core Workloads. (2021). https:\/\/github.com\/brianfrankcooper\/YCSB\/wiki\/Core-Workloads.  Brian F. Cooper. 2021. YCSB Core Workloads. (2021). https:\/\/github.com\/brianfrankcooper\/YCSB\/wiki\/Core-Workloads."},{"key":"e_1_2_1_15_1","doi-asserted-by":"publisher","DOI":"10.1145\/1807128.1807152"},{"key":"e_1_2_1_16_1","doi-asserted-by":"publisher","DOI":"10.5555\/2387880.2387905"},{"key":"e_1_2_1_17_1","volume-title":"14th USENIX Symposium on Operating Systems Design and Implementation (OSDI\u201920)","author":"Dai Yifan","year":"2020"},{"key":"e_1_2_1_18_1","doi-asserted-by":"publisher","DOI":"10.1145\/3318464.3389711"},{"key":"e_1_2_1_19_1","doi-asserted-by":"publisher","DOI":"10.5555\/2616448.2616486"},{"key":"e_1_2_1_20_1","doi-asserted-by":"publisher","DOI":"10.1145\/2815400.2815425"},{"key":"e_1_2_1_21_1","doi-asserted-by":"publisher","DOI":"10.1145\/3299869.3319860"},{"key":"e_1_2_1_22_1","doi-asserted-by":"publisher","DOI":"10.5555\/1316689.1316748"},{"key":"e_1_2_1_23_1","doi-asserted-by":"publisher","DOI":"10.1145\/2934872.2934908"},{"key":"e_1_2_1_24_1","doi-asserted-by":"publisher","DOI":"10.5555\/2946645.3007062"},{"key":"e_1_2_1_25_1","unstructured":"High-Performance Big Data (HiBD). 2021. RDMA-based Memcached (RDMA-Memcached). (2021). http:\/\/hibd.cse.ohio-state.edu.  High-Performance Big Data (HiBD). 2021. RDMA-based Memcached (RDMA-Memcached). (2021). http:\/\/hibd.cse.ohio-state.edu."},{"key":"e_1_2_1_26_1","volume-title":"et\u00a0al","author":"Jonas Eric","year":"2019"},{"key":"e_1_2_1_27_1","doi-asserted-by":"publisher","DOI":"10.5555\/3323234.3323236"},{"key":"e_1_2_1_28_1","doi-asserted-by":"publisher","DOI":"10.1145\/2740070.2626299"},{"key":"e_1_2_1_29_1","doi-asserted-by":"publisher","DOI":"10.5555\/3026877.3026892"},{"key":"e_1_2_1_30_1","doi-asserted-by":"publisher","DOI":"10.5555\/3026959.3027000"},{"key":"e_1_2_1_31_1","doi-asserted-by":"publisher","DOI":"10.5555\/3291168.3291200"},{"key":"e_1_2_1_32_1","doi-asserted-by":"publisher","DOI":"10.1145\/3183713.3196909"},{"key":"e_1_2_1_33_1","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359628"},{"key":"e_1_2_1_34_1","doi-asserted-by":"publisher","DOI":"10.1145\/3132747.3132756"},{"key":"e_1_2_1_35_1","doi-asserted-by":"publisher","DOI":"10.5555\/2616448.2616488"},{"key":"e_1_2_1_36_1","doi-asserted-by":"publisher","DOI":"10.1145\/2168836.2168855"},{"key":"e_1_2_1_37_1","doi-asserted-by":"publisher","DOI":"10.5555\/2535461.2535475"},{"key":"e_1_2_1_38_1","doi-asserted-by":"publisher","DOI":"10.5555\/3026959.3027001"},{"key":"e_1_2_1_39_1","doi-asserted-by":"publisher","DOI":"10.1145\/3318464.3380579"},{"key":"e_1_2_1_40_1","doi-asserted-by":"publisher","DOI":"10.5555\/2482626.2482663"},{"key":"e_1_2_1_41_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jalgor.2003.12.002"},{"key":"e_1_2_1_42_1","doi-asserted-by":"publisher","DOI":"10.5555\/3454287.3455008"},{"key":"e_1_2_1_43_1","doi-asserted-by":"publisher","DOI":"10.1214\/aoms\/1177729586"},{"key":"e_1_2_1_44_1","doi-asserted-by":"publisher","DOI":"10.1145\/3299869.3300069"},{"key":"e_1_2_1_45_1","doi-asserted-by":"publisher","DOI":"10.5555\/3026877.3026902"},{"key":"e_1_2_1_46_1","doi-asserted-by":"publisher","DOI":"10.1145\/3342195.3387519"},{"key":"e_1_2_1_47_1","doi-asserted-by":"publisher","DOI":"10.14778\/2311906.2311915"},{"key":"e_1_2_1_48_1","doi-asserted-by":"publisher","DOI":"10.14778\/3407790.3407836"},{"key":"e_1_2_1_49_1","doi-asserted-by":"publisher","DOI":"10.1145\/3064176.3064189"},{"key":"e_1_2_1_50_1","doi-asserted-by":"publisher","DOI":"10.1145\/3332466.3374547"},{"key":"e_1_2_1_51_1","unstructured":"The Transaction Processing Council. TPC-C Benchmark V5.11. (????). http:\/\/www.tpc.org\/tpcc\/.  The Transaction Processing Council. TPC-C Benchmark V5.11. (????). http:\/\/www.tpc.org\/tpcc\/."},{"key":"e_1_2_1_52_1","doi-asserted-by":"publisher","DOI":"10.1145\/3132747.3132762"},{"key":"e_1_2_1_53_1","doi-asserted-by":"publisher","DOI":"10.1145\/3035918.3056101"},{"key":"e_1_2_1_54_1","doi-asserted-by":"publisher","DOI":"10.1145\/2670979.2671002"},{"key":"e_1_2_1_55_1","doi-asserted-by":"publisher","DOI":"10.1145\/3409963.3410496"},{"key":"e_1_2_1_56_1","doi-asserted-by":"publisher","DOI":"10.1145\/2592798.2592815"},{"key":"e_1_2_1_57_1","doi-asserted-by":"publisher","DOI":"10.5555\/3291168.3291186"},{"key":"e_1_2_1_58_1","doi-asserted-by":"publisher","DOI":"10.1145\/2815400.2815419"},{"key":"e_1_2_1_59_1","doi-asserted-by":"publisher","DOI":"10.5555\/3358807.3358869"},{"key":"e_1_2_1_60_1","doi-asserted-by":"publisher","DOI":"10.5555\/3294996.3295058"},{"key":"e_1_2_1_61_1","doi-asserted-by":"publisher","DOI":"10.14778\/3055330.3055335"},{"key":"e_1_2_1_62_1","doi-asserted-by":"publisher","DOI":"10.1145\/2882903.2915222"},{"key":"e_1_2_1_63_1","doi-asserted-by":"publisher","DOI":"10.1145\/3299869.3300081"}],"container-title":["ACM Transactions on Storage"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3468520","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3468520","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,17]],"date-time":"2025-06-17T20:17:18Z","timestamp":1750191438000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3468520"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,8,16]]},"references-count":63,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2021,8,31]]}},"alternative-id":["10.1145\/3468520"],"URL":"https:\/\/doi.org\/10.1145\/3468520","relation":{},"ISSN":["1553-3077","1553-3093"],"issn-type":[{"value":"1553-3077","type":"print"},{"value":"1553-3093","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,8,16]]},"assertion":[{"value":"2021-01-01","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-05-01","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-08-16","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}