ALEX

Introduction

ALEX is a new class of learned indexes which addresses issues that arise when implementing dynamic, updatable learned indexes. Compared to the learned index from Kraska et al., ALEX has up to 3000× lower space requirements, but has up to 2.7× higher search performance on static workloads. Compared to a B+Tree, ALEX achieves up to 3.5× and 3.3× higher performance on static and some dynamic workloads, respectively, with up to 5 orders of magnitude smaller index size. Our detailed experiments show that ALEX presents a key step towards making learned indexes practical for a broader class of database workloads with dynamic updates.

Highlight

Publications: SIGMOD 2020 (research)

Collaborators:

MIT: Jialin Ding, Tim Kraska

Microsoft Research: Umar Farooq Minhas, David Lomet, Jae Young Do, Yinan Li, Chi Wang, Badrish Chandramouli, Johannes Gehrke, Donald Kossmann

ETH: Hantian Zhang

Jia Yu
Jia Yu
Assistant Professor (from Fall 2020)

Jia Yu obtained his PhD from Arizona State University in Summer 2020. His research interests include database systems, distributed data systems and geospatial data management.

Related