The purpose of this course is to teach operating system design from a research point of view. We will go over key topics of operating systems, walk through the evolution of many different operating systems techniques, and examine their usage in both important historical systems and in modern systems.
We will be going over the following topics:
This is a research-oriented seminar course with a major research project. In short, the course is about discussing and learning (by doing!) computer systems research.
Prerequisite: CS 423, ECE 391, CS 425, and CS 433 (or equivalent). Take the Prerequisite Quiz if you are not sure.
The course does not have a textbook. Instead, the course material will come from seminal, noteworthy, or representative research papers and articles. Each lecture (except the first one) will have two assigned papers to read. You are expected to read these papers before coming to class, and be prepared to discuss them. We will also list recommended readings; you are encouraged to read those, but not required.
I highly recommend you to read Griswold's advices on how to read a research paper. The take-home message is that until you can answer a bunch of questions, you are not done reading a paper.
I also strongly encourage you to discuss the papers with other students in the class — you may have insights that others do not, and vice versa. Oftentimes, students form reading groups, which I heartily encourage; on the other hand, I would like to point out that group discussion is not an effective substitute for actually reading the paper.
If you need a textbook to review and catch background, I recommend the OSTEP book.
The following table outlines the schedule for the course. We will update it as the semester progresses.
| Date | Topic | Paper | Project | Guest |
|---|---|---|---|---|
| 1/22 | Course Intro (slides) | N/A | Brainstorming and teaming up | TBD |
| 1/24 | Historical Perspectives (1) | THP, Nucleus | Brainstorming and teaming up | TBD |
| 1/29 | Historical Perspectives (2) | TENEX, HYDRA | Brainstorming and teaming up | TBD |
| 1/31 | Historical Perspectives (3) | UNIX, Plan 9 | Brainstorming and teaming up | TBD |
| 2/5 | Kernel Architecture (1) | L4 (qual paper) | Brainstorming and teaming up | TBD |
| 2/7 | Kernel Architecture (2) | Exokernel | Proposal due: 2/9 11:59pm | TBD |
| 2/12 | Synchronization (1) | Mesa | Presentation and discussion | TBD |
| 2/14 | Synchronization (2) | Chess | Presentation and discussion | TBD |
| 2/19 | Scheduling (1) | TBD | Presentation and discussion | TBD |
| 2/21 | Scheduling (2) | TBD | Presentation and discussion | TBD |
| 2/26 | Memory Management (1) | Mach MM (qual paper) | Presentation and discussion | TBD |
| 2/28 | Memory Management (2) | ECPT | Presentation and discussion | TBD |
| 3/5 | File Systems (1) | LFS (qual paper) | Presentation and discussion | TBD |
| 3/7 | File Systems (2) | Rio | Checkpoint #1 due: 3/9 11:59pm | TBD |
| 3/12 | Network Stack (1) | TBD | Presentation and discussion | TBD |
| 3/14 | Network Stack (2) | TBD | Presentation and discussion | TBD |
| Fall Break | Food. Lots of it. | Sleep. Lots of it. | N/A | |
| 3/26 | Virtual Machine (1) | Xen (qual paper) | Presentation and discussion | TBD |
| 3/28 | Virtual Machine (2) | LightVM | Presentation and discussion | TBD |
| 4/2 | Container (1) | TBD | Presentation and discussion | TBD |
| 4/4 | Container (2) | TBD | Checkpoint #2 due: 4/6 11:59pm | TBD |
| 4/9 | Extensibility (1) | SPIN | Presentation and discussion | TBD |
| 4/11 | Extensibility (2) | TBD | Presentation and discussion | TBD |
| 4/16 | Reliability (1) | Bugs as deviant behavior | Presentation and discussion | TBD |
| 4/18 | Reliability (2) | Nooks (qual paper) | Presentation and discussion | TBD |
| 4/22 | Security (1) | KCFI | Presentation and discussion | TBD |
| 4/25 | Security (2) | Haven | Presentation and discussion | TBD |
| 4/30 | D.i.s.tri...buuuuuted | MapReduce | Presentation and discussion | TBD |
| 5/2 | Machine Learning | TBD | Presentation and discussion | TBD |
| 5/7 | Future | Something wild | Final report due 5/7 11:59pm | TBD |
Since this is a research-based course, class participation is required. We (everyone) will discuss the papers and our research progress in class. Note that your performance in class, including presentation and discussion form 45% of your overall grade, so it does matter that you both show up to class and participate in the discussion.
The best way to learn is by doing. You will undertake your own research project individually or in a group of two, or three if it is a large project. A group with a size larger than two is not encouraged, but is possible if you have a strong justification that the project needs more members. I will provide a list of ideas to get you started thinking, but I highly encourage you to pursue your own ideas which typically lead to better results. You will write a project report and present it at the end of the course. The details of the research project is described in the following link.
You can find a list of 523 projects that were subsequently published in workshops, conferences, and journals.
We do not release information that are not opened to public to protect students' work (many of which are closely connected to their thesis research).
Note: You are expected to be aware of Academic Integrity Guidelines of the University of Illinois. Any violation of the course or university policies will be treated seriously, and could lead to severe repercussions. Pleae don't cheat. It's not worth it.
There is no homework, no midterm, no final exams. The course is about discussing Systems research (10%) and doing Systems research (90%). Note that the Project Proposal, Checkpoint Report 1, and Checkpoint Report 2 account for 5%, 5%, and 5%, respectively.
This course is designed based on CSE 221 (Graduate Operating Systems) taught at University of California at San Diego (where I did my PhD). It was my favourite course during my grad school -- I took it and TA-ed for it multiple times with different professors including Geoff Voelker, Yuanyuan Zhou, and Stefan Savage. They should take the credits for the course design.