Students will study advanced operating system topics and be exposed to recent developments in operating systems research. This course involves readings and lectures on classic and new papers. Topics: operating system design, virtual memory management, virtual machines, OS interaction with the hardware architecture, synchronization and communication, file systems, protection, and security.
Assignments, announcements, and schedules will appear on either the web page, or I might email.
http://www.cs.utexas.edu/~witchel/380L
The prerequisite for this class is CS372 or the equivalent. It is necessary to have this background before taking the class, as we'll read a lot papers quickly without much time for catching up on the basics. The course assumes an understanding of topics in operating systems such as synchronization, virtual memory management, scheduling, and file systems.
The course consists of readings, homeworks, discussion, in-class presentation, two exams, and an independent research project (and possibly pop quizzes). The two most important things to know about the class: (1) the main goal is to have interesting in-class discussions and (2) we recommend you read each paper at least twice, preferably more than a day in advance so that it sinks in.
Most of the work in this
course consists of reading journal and
conference papers. Most of the papers we read will be good,
and all
are influential. We will cover one to two papers at each class
meeting.
This class will be primarily discussion based. Active discussion will
(hopefully)
give you a non-trivial understanding of the material.
This class is larger than ideal for a discussion-based class. I believe we can overcome this hurdle if everyone comes to class well prepared to participate in (or lead) discussions. If you skim papers and space out during discussion, you are doing your colleagues a disservice by lowering the level of discussion. The other problem, of course, is that individually you will learn much less if you follow that approach. The assignments and grading, therefore, are focused on creating an atmosphere where everyone comes to class well prepared for discussion.
Class time will not be used to rehash the material in the papers. Instead, it will be used to highlight the important points and discuss some of the more interesting features. There will be as much as 10-15 hours of reading per week. Do not take this course unless you are willing and able to do a lot of reading.
Paper
critiques. We will read about 28 "core technical" papers in class
(plus a number of "background" and "professional development" papers.)
For each of the core technical papers, you will prepare a
(approximately) 1-page written critique (here is
a template). This critique must be typeset (no handwritten
critiques will be accepted) and must be turned in before the
start of the class that discusses the paper. You may skip up to
4 critiques with no grading penalty. Details of what we expect in a
critique will be covered in a handout.
Class participation. The class will provide ample opportunity to get involved in the discussions, and if you do not regularly participate in discussions, you will not get as much out of the class as you could. Your class participation will be based on three factors: (1) the TA and instructor's assessment of whether you were regularly involved in the discussions over the course of the semester, (2) turning in acceptable critiques as described above, (3) the in-class presentation of at least one paper (4) there might be several pop quizzes on the reading assignments over the course of the semester. If you turn in a critique (e.g., claim to have read and understood the paper) but your quiz indicates that you don't understand the paper, we will lower your class participation grade. (Note that if you don't turn in a critique for a class, you are "exempt" from taking a quiz for that class.)
Readings. There is no textbook for this course. The course
is based on a collection of journal and conference papers that
describe the history and state of the art in operating systems. The
preliminary list of papers and schedule is available on-line. If you are interested, more papers
on these topics are available from this reading list ; we will cover a subset of these
papers, and we will cover topics roughly in the order listed
there. You must read the papers before class. At a minimum we
recommend two close readings. We will provide most papers online;
those that are only available in hardcopy will be provided about a
week before they are needed.
Written homeworks. We may assign additional written homeworks that cover basic experimental skills useful for studying operating systems (cache simulation, network simulation, and simple statistics and data analysis). Unless otherwise noted, the homeworks will be done individually.
To test your understanding of the material, there will be two midterm exams. I plan to conduct them in class, but I may change one or both to be take-home exams.
Special offer: you can write your own exam questions! Submit a question with your solution in advance of the exam, and if we like it, it will appear on the exam.
This course requires several programming assignments that will give
you
experience in building, booting and running an operating system.
The
assignments will also expose you to methodological systems issues such
as how to model, measure and report performance, how to design a
workload to test kernel functionality, and the dependence on workload
for the evaluation of a system feature. Finally the assignments
will
expose you to how to write about systems, their design, implementation,
and measurement.
These assignments should demystify the operating system, convincing
you that the OS really is just a program. Sometimes puzzling
system
behavior can be understood and worked around by reading and
understanding the source code of the OS. Why did mmap return
ENOMEM?
There are several distinct possibilities that you can see in the
code.
These assignments might even give you a bit of practical knowledge, for
example allowing you to get Linux to recognize your fancy, new USB
device.
I strongly encourage you to discuss the papers and the homeworks with anyone you can. That's the way good science happens. As a professional, you should acknowledge significant contributions or collaborations in your written or spoken presentations.
The paper critiques should reflect your understanding of the paper. It is not acceptable to turn in a summary if you have not made an honest effort to read the paper. If you don't have time to read a paper before a class, make use of one of your skip credits. Never read another student's summary before you have turned in your own.
Unless otherwise stated, the homeworks and programming assignments must be done individually. You may orally discuss the homeworks with anyone, but you may not look at anyone else's code and you may not allow anyone else to look at your code.
Exams are to be done
individually. They may only be discussed with
the instructor and TA.
Intellectual dishonesty can end
your career, and it is your
responsibility to stay on the right side of the line. If you are not
sure about something, ask.
Please read the department's code of conduct.