Arctic: A Practical Lock-Free Adaptive Radix Tree

Newton Ni, Nicolas Garza, Jenny Stinehour, and Michael Goppert, The University of Texas at Austin; Michal Friedman, ETH Zürich; Emmett Witchel, The University of Texas at Austin

Indexing data structures are vital to the modern systems ecosystem, but there are no indexes that offer high performance, lock freedom, and range scans. Arctic is a lock-free adaptive radix tree that achieves all three: Arctic outcompetes lock-based indexes, including a concurrent hash map, on many YCSB configurations, guarantees non-blocking operation through careful metadata layout and an (eponymous) freezing-based coordination protocol, and offers non-linearizable range and prefix scans. Arctic also contributes a novel safe memory reclamation scheme that uses operation keys to approximate reachable pointers. We integrate Arctic into RocksDB and Turso, improving throughput up to 40% and 12% on their write-heavy benchmarks relative to their default skiplist indexes.

OSDI '26 Open Access Sponsored by
King Abdullah University of Science and Technology (KAUST)

Open Access Media

USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.

BibTeX
@inproceedings {318461,
author = {Newton Ni and Nicolas Garza and Jenny Stinehour and Michael Goppert and Michal Friedman and Emmett Witchel},
title = {Arctic: A Practical {Lock-Free} Adaptive Radix Tree},
booktitle = {20th USENIX Symposium on Operating Systems Design and Implementation (OSDI 26)},
year = {2026},
isbn = {978-1-939133-55-7},
address = {Seattle, WA},
pages = {739--754},
url = {https://www.usenix.org/conference/osdi26/presentation/ni},
publisher = {USENIX Association},
month = jul
}