Don't spend time regurgitating obvious stuff from the papers! The point of these questions is to
and then to write down what it is that you thought about. I read 30-odd write-ups, so the more interesting you are, the better! If you find yourself just repeating a lot of details from the paper, you are going down the wrong path.
Don't spend time criticizing how they wrote the paper. That is, I don't need to know whether you thought the paper was well written or not. We'll have plenty of time to talk about those types of things later in the class...
No more questions: you are all done with that! Just please read the papers. Eventually, I will even catch up with the grading of all your wonderful write-ups. I promise! Now please go work on your projects.
Question for the week (11/25)
Q14, due Friday, 11/30:
Do distributed systems with centralized components scale? Use the readings to guide your answer.
Question for the week (11/18)
Q13, due Wednesday, 11/21:
NFS and AFS are two classic distributed file systems. Describe the strengths and weakness of each file system.
Question for the week (11/11)
Q12, due Friday 11/16:
The Locus paper presents a system that allows you to run jobs on remote machines. The Map-Reduce paper presents a system that allows you to run jobs on remote machines. Yet Map-Reduce is quite successful, while (pretty much) no one is using a system like Locus. Why do you think that is so?
Question for the week (10/28)
We are now moving to one question per week, to give you more time for your projects. Congratulations! The question will be due on
Multics and VMS
Multics is the root of all good ideas in modern computer systems, or so some say. Name and describe two good ideas that you see in the Multics design that have influenced modern computer systems.
Questions for the week (10/21)
No questions this week -- time to start getting ready for the exam!
Questions for the week (10/14)
Monitors and Mesa
Q10, 10/17: In theory, there's no difference between theory and practice, but in practice, there is. Discuss how the "practice" as described in the Mesa paper differs from the "theory" described in the Monitors paper.
Q11, 10/19: Describe how lottery scheduling can be used for managing disk performance. What makes this harder than using lottery for CPUs?
Questions for the week (10/07)
Q8, 10/9: As we saw in this paper, sometimes waiting can make things run faster. Describe one or two other cases in operating systems where waiting to do something can actually improve performance.
Q9, 10/11: RAID protects from certain types of disk faults. In the IRON FS work, we heard about a large number of disk faults. What types of disk faults might standard RAID systems have trouble dealing with?
Questions for the week (09/30)
Q6, 10/2: The fsync() system call forces writes to disk immediately. Imagine a program that issues each write and then always calls fsync(). How would such a program perform on LFS?
Q7, 10/4: The IRON FS paper is probably one of the best things you'll ever read. Describe three things about IRON FS which you absolutely love.
Questions for the week (09/24)
Q5, 09/26: What are three main weaknesses with Nooks? (note: you must bring up at least one with Professor Swift on Wednesday in class)
Q6, 09/28: Write a short program (in pseudocode, please!) that performs terribly on FFS. Assume you have only the ability to access existing files (no file creations are allowed).
Questions for the week (09/17)
Q3, 09/19: How does exokernel differ from Nucleus? How is it the same?
Q4, 09/21: Criticize the evaluation of Disco. What is believable, what is suspect?
THE and Nucleus
Q1, 09/12: THE and Nucleus are old systems, but in many ways were influential. Which aspects of these classics do you see in modern operating systems?
Q2, 09/14: Systems are all about assumptions. What assumptions do the Pilot authors make about PCs that were good, and what assumptions do they make that were bad?