Research

I have done some research. My research interests are mainly in computer architecture, operating systems, and security. More specifically, I am currently most interested in trusted execution environments (TEEs), sandboxing and isolation, and heterogeneous computing.

Publications

Adaptive and Efficient Dynamic Memory Management for Hardware Enclaves

Vijay Dhanraj, Harpreet Singh Chawla, Daniel Manila, Eric Thomas Schneider, Erica Fu, Mona Vij, Chia-Che Tsai, Donald E. Porter

ACM International Systems and Storage Conference (SYSTOR), 2025

Mechanisms of Memory Protection

Eric Thomas Schneider

MS Project

Service

Research Projects

  • Adapting sandboxing to TEEs; this project is my current focus. Inspired by papers like Ryoan, I am exploring a powerful threat model where neither an application nor a hypervisor is not trusted to not leak sensitive data. We are exploring covert channels between an application and hypervisor and defenses for them, particularly for large language models. I am working on a paper on this topic.
  • "SATerm" - eBPF Termination. Under Professor Dan Williams (ROSA), I contributed to a project exploring the termination of long-running eBPF extensions. We submitted to USENIX ATC '25 with mixed reviews; currently the paper is being lead by another student, and some of the work is being upstreamed to Linux.
  • Media Attestation. This project is being lead by Jesse Wei. Under Professor Andrew Kwong, we are exploring using ARM CCA and the C2PA standard to prove the authenticity of images generated by a mobile camera.
  • "Native Enclave" - for our Research in Security class project, I worked with Yulu Pan on porting Keystone, a RISC-V TEE implementation, to uClinux (Linux without virtual memory), to try to explore how the idea of TEEs could be extended to become a first-class memory protection mechanism in a "native enclave" system. Unfortunately, Keystone is pretty tied to virtual memory, but I hope to revisit this idea for a future project.
  • ADΔER. Under Andrew Freeman, Professor Ketan Mayer-Patel, and Professor Montek Singh, I contributed towards the ADΔER asyncronous video research software suite. In Rust, I contributed a transcoder bandwidth limiting feature, implemented an event logging feature for the Rust implementation of the DVS Fast algorithm, fixed a couple Windows-related ADΔER issues, and updated the CI (with GitHub Actions). I also did a survey paper. In summer 2023, I was funded by the Laboratory for Analytical Sciences (LAS) for this work.
  • NeuroRuler. Under Professor Martin Styner (NIRAL), with a few other students, I helped create this research software, written in Python. Basically, using MRI data, it will calculate head circumference given a particular "slice" of the brain. The software has some options for rotation, smoothing, threshold, and a batch API via the command line. It is being used by some neuro imaging researchers in Martin's lab.
  • Gramine - Improving SGX Performance with Cooperative, Dynamic Page Management. Under Professor Donald Porter (OSCAR), I contributed to this paper mainly with writing and benchmarking.