End of Quarter Work

General

Each individual knows best what the best use of their time is over the next week. We trust your judgement. At the same time, we urge you to push yourself just a bit more on classwork and other investments in your future than you might feel you want to. That said:

• There is no final exam.
• Lab 4 and the final exercise are optional.
• If you do them it cannot lower your grade relative to the grade you'd get if you hadn't done them. Grading will be done normally, as if times were normal. If your grade is above your average through Lab 3, the new grade raises it. If your grade is below your average through Lab 3, it is not lowered.
• There is a symbolic reward for each of the two non-mandatory items that remain. (That means there will be some credit for doing each, but the credit will be worth way under the amount required to change the final grade by a 10th of a point -- .03, for instance.)
• There are two options for the final exercise.

For what it's worth, I feel that if you take seriously what is being asked, that you take responsibility for deciding what's best for you and then carry that out, that you should take pride in doing so no matter what your decision is.

Lab 4

• Completion of Lab 4 is optional, as explained above. There is no penalty for not handing it in.
• The due date has been extended through the weekend.
• Grading will be in-person, so the extension is really until your grading session.
• If your partner decides not to continue working on it, you can still schedule a grading session and receive credit for what you have done, even if the project isn't working.

Final Exam Activity

• It's optional, as explained above.
• If you do it, please hand in your work by Saturday morning, June 13.
• I will read and grade all submissions.
• There are two options to choose between.
• What to do for each is explained below.
• Turn-in is on Canvas.
• I encourage you to at least read the abstracts (first few paragraphs) of each paper cited by Option 1 whether you intend to do that exercise or not (including if you intend to choose neither option).

Option 1

I have selected three papers from the most recent Symposium on Operating Systems Principles (SOSP). SOSP is one of the top research conferences in operating systems. It is very competitive to get a paper accepted to it.

The exercise is:

• Select one of the papers and read it.
• Spend some reasonable effort on it. It is expected that it will be challenging to read - the intended audience is researchers in operating systems. I'm hoping that you will learn things from reading the paper, and that takes some work. At the same time, I'm not expecting you to feel that you understand everything the paper says. (That's usually challenging for anyone, including the target audience.) Spend "reasonable time" on it, and stop when you've reached the point of diminishing returns in what you're gaining per 10 minutes spent.
• Create and hand in "appropriate length" report that addresses at least the following:
  • Your name
  • Title of the paper you're writing about
  • What is the paper about? What issue does it address? Why is that issue important?
  • What is the paper's contribution to knowledge related to its issue? How might the world be changed by the paper's results?
  • What topics from CSE 451 this quarter were directly useful to you when reading the paper, and how were they relevant?
  • What did you learn from the paper? How did it go beyond what we talked about this quarter?
  • If there were aspects of the paper you ended up unclear about, what were they? Were there such large gaps between the background you got from 451 and the place the paper started from that you felt unsure if you understood what it was saying? If so, can you describe what you think was missing from your background knowledge?
• Short and straight to the heart of the matter, rather than exhaustive detail, is always better. "e=mc²" rather than "If a body gives off the energy L in the form of radiation, its mass diminishes by L/V². This equation is approximated by neglecting magnitudes of fourth and higher orders of a series expansion." (Thanks, Wikipedia.)

Here are the papers. I've ordered them from what I guess is least challenging to most challenging. (To see the full paper, click on the big red PDF buttons on the pages these links take you to.)

• An Analysis of Performance Evolution of Linux's Core Operations
• Scalable and practical locking with shuffling
• File Systems Unfit as Distributed Storage Backends: Lessons from 10 Years of Ceph Evolution

Option 2

IF you have have created a detailed study gude for this course, you can:

• Submit it. (You still own it. We won't use it for anything.)
• Submit a 1-page "most critical points" summary of it.

The one page limit is not intended to make this almost no work, it's intended to make it hard. You have to distill what you have down to a single page (without using font size tricks). Presumably that will require thinking about all the content and finding some way to express only the most important points.