We present XIndex, a concurrent ordered index designed for fast queries. Similar to a recent proposal of the learned index, XIndex uses learned models to optimize index efficiency. Comparing with the learned index, XIndex is able to effectively handle concurrent writes without affecting the query performance by leveraging fine-grained synchronization and a new compaction scheme, Two-Phase Compaction. Furthermore, XIndex adapts its structure according to runtime workload characteristics to support dynamic workload. We demonstrate the advantages of XIndex with both YCSB and TPC-C (KV), a TPC-C variant for key-value stores. XIndex achieves up to 3.2× and 4.4× performance improvement comparing with Masstree and Wormhole, respectively, on a 24-core machine, and it is open-sourced.
Tue 25 FebDisplayed time zone: Tijuana, Baja California change
15:45 - 17:00 | Search and Index (Mediterranean Ballroom)Main Conference Chair(s): Idit Keidar Technion - Israel institute of technology | ||
15:45 25mTalk | Non-Blocking Interpolation Search Trees with Doubly-Logarithmic Running Time Main Conference | ||
16:10 25mTalk | YewPar: Skeletons for Exact Combinatorial Search Main Conference Blair Archibald University of Glasgow, Patrick Maier University of Stirling, Rob Stewart Heriot-Watt University, Phil Trinder University of Glasgow | ||
16:35 25mTalk | XIndex: A Scalable Learned Index for Multicore Data Storage Main Conference Chuzhe Tang Shanghai Jiao Tong University, Youyun Wang Shanghai Jiao Tong University, Gansen Hu Shanghai Jiao Tong University, Zhiyuan Dong Shanghai Jiao Tong University, Zhaoguo Wang Shanghai Jiao Tong University, Minjie Wang New York University, Haibo Chen Shanghai Jiao Tong University | ||