CSE-624: Advanced Operating Systems, Spring 2005

Important notes about this class:

• The focus of the class in S'05 will be on operating systems projects, especially file systems and security projects (but other topics such as Networking are also possible).

• Project groups can vary from a single person to as many as 4–5, as long as the project's scope matches the group's size.

• Each group of students will pick a suitable project at the beginning of the class: either a new project, or an on-going one to be extended. If you don't have a project in mind, don't worry; you can join one of the existing teams in this class, or we'll suggest something to you. Group members will present their project idea, design, progress reports, and final demo in front of the whole class.

• In the beginning of the semester, before we get too deep into projects, we will be reading select papers from the literature and discuss them in class. We will pick a number of papers representing a range of papers from classics (original Multics papers) to the latest cutting-edge ones (e.g., "The Google Filesystem"). Each student will be expected to lead the discussion of 1–2 papers during the semester.

• I (with the help of my students) will be giving specific focused lectures intended to discuss a specific issue such as: VFS internals, NFSv4, kernel debugging techniques, file system benchmarking, etc.

• One of the main goals of the class is to train graduate students for conducting research, writing and publishing, and presenting papers. I will give lectures showing how to prepare a good set of slides for use with a computer projector and PowerPoint/PDF; how to give a talk; how to practice talks; how to write a short 200-word abstract for an idea; how to write a one-page summary paper, a 4–5 page "extended abstract" paper, or a full-length 14–16 page conference paper; I will give lectures covering the common mistakes made by authors/speakers, and more. By the end of this class, you should feel comfortable with all of the stages of research: coming up with an idea, defending your idea, conducting the actual research, writing it up, presenting it orally, and scientifically criticizing your own and others' work.

• For this class to be successful, it is important that you attend it each week and participate in the open discussions.

• Grading in the class will consist of several components. The most important component will be your project: presentations and writings; in particular, by the end of the semester you will be expected to have produced a professional quality conference-style 12–14 page research paper. What I will look for in particular is how much you have improved your research and presenation skills over the semester (I will not grade on a curve). (It is my hope that you could submit that paper to a future conference after the semester is over.) Other grading components I will take into account include your presentation and leading of the discussion of the 1–2 research papers, and your overall participation and attendance in class. There will be no exams.

• To take this class, you should (1) have taken an introductory course in operating systems before so you know all the basics, and (2) have taken before and passed a course such as CSE-506, CSE-508, or CSE-509 (i.e., a course with a substantial "systems" programming project component). Alternatively, you should have proven experience from elsewhere (perhaps where you worked before) that you can conduct phd-level research. Note that good English language skills are not required to take this class: the whole point of this class is that I want to help you to improve those skills, no matter what your existing skills or native language happen to be!