Research Project
PhD students or students interested in research may elect to do a research project instead of Projects 3 and 4. Ideally, this is a project you would continue to work on past the class, and possibly publish. Please submit a short (2 paragraphs) project proposal by Friday, March 18 if you plan to do this. Send to me over e-mail. You may choose your own research project (pending my approval), but here are some ideas:
- Validate experiments from LegoOS paper. Explore other workloads. Port it to another language, evaluate it on different hardware, explore other hardware devices, e.g., FPGAs.
- Explore Theseus OS, analyze its performance in a virtualized environment.
- Explore replication for fault tolerance in the context of virtual subroutines (virtines). See our paper here.
- Add language support to WASM to support virtines.
- Apply program synthesis to virtine execution environments.
- Use modern OS and virtualization techniques (e.g., Theseus) to revisit microreboot
- Explore usage of high-performance I/O in serverless functions. Maybe consider adapting Demikernel architecture.
- Extend/improve J-NVM
- Improve stateful serverless computing over shared logs with NVM
- Implement an MPK-based back-end for virtines
- Work on/improve/extend a system from VEE ‘21