Seminar overview

In this seminar we will be discussing the state-of-the-art of systems research. This is a great way to learn technical matters but also how to do systems research!

This semester we are going to mostly focus on papers from OSDI, SOSP, EuroSys, and ASPLOS, where some of the best systems, systems security, and architecture work is published, so there is always a lot to learn from these papers.

We meet once a week for one hour and have an informal discussion about the papers – this is not a course, and you will not get credits for it, but you will probably learn a lot.

Make sure you subscribe to the seminar mailing list to ensure that you get notifications about paper discussions.

Guidelines for everyone

To participate in the seminar there are a couple of expectations for every participant that help ensure we have lively and informative discussions:

• ATTENDANCE Attend (almost) all seminar meetings (meetings are on Monday @ 10:30am-11:30am).
• READING Read the papers beforehand; typically 1 paper per meeting.
• PARTICIPATION Actively participate in the seminar discussion and lead the discussion of a couple of papers.

Guidelines for discussion leads

As a discussion lead your primary goal is to foster and manage the paper discussion. This requires understanding well the material, preparing some slides, managing time, and managing the discussion during the seminar. Please try to follow the following guidelines when leading the discussion:

• SLIDES Avoid using the author slides. Author slides tend to be about selling their own work and tend to sweep under the rug the limitations. In addition, each presenter has his own presentation style so it is never ideal to use someone else's slides.
• STRUCTURE You are free to decide how to structure the talk, but you are encouraged to prepare 15-20 slides. Avoid spending more than 20-30 minutes talking so that there is enough time left for discussion. In addition, it is important to prepare questions and discussion points ahead of time to foster interesting discussions. Your short presentation should clearly highlight: (a) which problem the authors are solving, (b) why this problem is important, (c) how they tried to solved the problem and how it relates to previous approaches, and (d) how well did they solve the problem.
• SETUP Please make sure to arrive early and test the screen casting ahead of time (e.g., 5-10 minutes) to avoid delays.

Tentative Schedule

Date	Paper	Lead
1/27	Resolving Code Review Comments with Machine Learning. ICSE ’24	Rui
2/3	Fast Core Scheduling with Userspace Process Abstraction.SOSP 24	Dinglan
2/10	SwiftPaxos: Fast Geo-Replicated State Machines. NSDI ’24	Congyu
2/17	verus: A Practical Foundation for Systems Verification, SOSP'24	Paul
2/24	Efficient Reproduction of Fault-Induced Failures in Distributed Systems with Feedback-Driven Fault Injection. SOSP'24	Ajay
3/3	Unearthing Semantic Checks for Cloud Infrastructure-as-Code Programs. SOSP'24.	Gemma
3/10	A Secure, Fast, and Resource-Efficient Serverless Platform with Function REWIND. ATC '24	Rahul
3/17	NO SEMINAR [SPRING BREAK]
3/24		Silver
3/31	Principled Microarchitectural Isolation on Cloud CPUs. CCS'24	Qi
4/7	When Threads Meet Interrupts: Effective Static Detection of Interrupt-Based Deadlocks in Linux. USENIX Security'24	Sishuai
4/14	Understanding the Linux Kernel, Visually. EuroSys'25	Chih-En
4/21		Rui
4/28		Dinglan