Occualizer: Optimistic Concurrent Search Trees From Sequential Code


Tomer Shanny and Adam Morrison, Tel Aviv University


This paper presents Occualizer, a mechanical source code transformation for adding scalable optimistic synchronization to a sequential search tree implementation. Occualizer injects synchronization only to the update steps of tree operations, leaving traversal steps to execute unsynchronized, thereby maximizing parallelism.

We use Occualizer to create concurrent versions of a sequential B+tree, trie, and red-black tree. Evaluation on a 28-core machine shows that Occualizer's trees significantly outperform prior mechanically-crafted trees on non-read-only workloads and are comparable (within 4%) on read-only workloads. Overall, Occualizer shrinks the performance gap between mechanically- and hand-crafted trees by up to 13×. When using Occualizer's B+tree as the index in the STO main-memory database, the system's throughput degrades by less than 30% compared to the default Masstree index, and it scales better.

OSDI '22 Open Access Sponsored by NetApp

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.

@inproceedings {280882,
author = {Tomer Shanny and Adam Morrison},
title = {Occualizer: Optimistic Concurrent Search Trees From Sequential Code},
booktitle = {16th USENIX Symposium on Operating Systems Design and Implementation (OSDI 22)},
year = {2022},
isbn = {978-1-939133-28-1},
address = {Carlsbad, CA},
pages = {321--337},
url = {https://www.usenix.org/conference/osdi22/presentation/shanny},
publisher = {USENIX Association},
month = jul

Presentation Video