Audrey Cheng
UC Berkeley
accheng@berkeley.edu
Bio
Audrey Cheng is a fifth-year Ph.D. student studying computer science at UC Berkeley where she is advised by Natacha Crooks and Ion Stoica. Her research centers around providing performance and correctness guarantees for large-scale, distributed data management systems. Her work is supported by a Meta Research PhD Fellowship, a NSF GRFP Fellowship, and a Berkeley Chancellor’s Fellowship and was the recipient of a Best Industry Paper Award at VLDB 2021. Audrey completed her B.S.E. at Princeton University, where she was advised by Wyatt Lloyd.
Areas of Research
- Computer Systems
Optimizing Performance for Modern Data Systems
Today, hyper-scale, data-intensive computing is more ubiquitous than ever. Modern services, such as the billion-user, geo-replicated applications offered by Google and Meta, require performance and reliability as well as safety and correctness guarantees. As a result, they have placed increasing demands on existing data management solutions and have led to a host of emerging research challenges. My research focuses on the core problems in these large-scale, distributed data management systems. In particular, maximizing throughput is crucial to high-performance database systems, which focus on minimizing data access conflicts to improve performance. However, finding efficient schedules that reduce conflicts remains an open problem. To address this, I propose systematically exploring the entire schedule space, proactively identifying efficient schedules, and executing them precisely during execution to improve throughput. Our scheduling policy efficiently finds fast schedules and outperforms state-of-the-art search techniques. Furthermore, I have worked on how to achieve transactional guarantees on a hyper-scale system (serving billions of queries per second) and optimizing memory management for transactional workloads. My research has been adopted by Meta as well as leading databases, including PlanetScale and TiDB.