Reverse engineering the AMD Translation Look-aside Buffer

Cyber security has increasingly become more of a concern throughout the world.

Attacks are becoming more sophisticated forcing the security community to respond accordingly. Side-channel attacks take advantage of computer systems rather than flaws in algorithm implementations. Producing strong defences against side-channel attacks requires an intricate understanding of the computer system being hardened. 

This project’s primary objective is to understand how a minor component of AMD processors work. The component being studied is known as the Translation Look-aside Buffer (TLB). It acts like cache, bridging the difference in processing and memory access speeds exclusively for translations from virtual to physical addresses. Understanding the internal mechanics of AMD TLBs is the first step in discovering plausible attacks and developing suitable corresponding defences.

Theme
Securing our future

Booth
SF40

School
Computer Science

Exhibitor
Jason Lam

vote for this project: sf40

Back to project list