JOIN VIRTUALLY VIA ZOOM: https://utexas.zoom.us/j/96785148012

Persistent memory (PM) is a new storage technology that combines the speed and byte-addressability of DRAM with the durability of hard drive and solid-state drives. Recent research has explored using PM in applications like databases and file systems to maximize both performance and reliability. For example, file systems designed for PM achieve better performance than file systems on traditional storage media, while simultaneously providing stronger and clearer consistency guarantees. However, building correct applications for PM is difficult: developers must reason about fine-grained PM management and use assembly instructions to flush and order updates to storage media. As a result, applications built for PM are prone to bugs that can cause data loss or corruption in the event of a crash. This talk will discuss my work on Chipmunk, a tool for testing PM file systems for crash-consistency bugs, and my current work on a PM file system with statically-checked crash-consistency guarantees.

Hayley Leblanc is a third year PhD student advised by Vijay Chidambaram in the UT Austin Computer Science department. Her research focuses on file and storage system reliability, and she is particularly interested in using techniques from formal methods to reason about the correctness of these systems.