Read Tanenbaum, Chapter 1. Re-read Ward and Halstead (6.004 book), Computation Structures, chapters 16 and 18. Our technical content will be launched from Ward and Halstead as a base, so make sure you understand it securely.
If you have been paying attention, you know that reading reports are required on Tuesdays, not Thursdays. On Thursday, just come prepared to carry on the suggested discussion. Note that if you don't do the reading, your recitation instructor will notice that your participation isn't up to that of everyone else. Your goal is to make a good impression; reading the paper is the first step in making that good impression.
In preparation of this section, read readings #9 and #10. The topic of this section is system structure. Reading #9 is an "old" paper (1970); it is one of the first papers discussing microkernel-based systems, although the word was not used back then. Reading #10 is an article that appeared in Byte magazine in 1994. This article has a high hype density and basically repeats the arguments of reading #9. Cutting through the hype, you may want to ask yourself the question who is really benefiting from the microkernel trend. End-users? Hardware designers? Hardware vendors? OS programmers and vendors? Application programmers?
We will attempt to hand back your corrected reading reports in this section, so you can use your TA's feedback to improve your next writing assignment.
There will be an additional section for students enrolled in the lab during special lecture hours (2-3p) in Room 36-144. Dawson Engler will explain the first lab project (Simple TCP) in more detail and answer any questions you might have.
Tuesday is a virtual Monday, so there will be a lecture instead of a section. Nevertheless, you are supposed to hand in your one-page reading report on Tuesday in lecture. The reading assigned is Tanenbaum Chapter 3. You can skip Section 3.3.3 (Examples of paging hardware) and Section 3.7.2 (Segmentation with paging: Multics). Your reading report should address the following question:
Is the extra functionality of virtual memory really worth the added complexity in light of recent trends towards increased RAM? Without VM processors can be simplified, allowing them to become smaller, faster, or offer additional functionality. Operating systems can be simplified allowing them to be faster and more reliable. However, if anything, modern systems have increasing support for virtual memory. If you were designing a system would you get rid of virtual memory or not? Explain your reasoning.
In preparation for lecture, read Tanenbaum Chapter 2. You can skip sections 2.2 (interprocess communication) and 2.3 (classical IPC problems), if you like; they will be assigned for next week.
System aphorism of the week
Everything should be made as
simple as possible, but not simpler. (A. Einstein)
6.033 Handout 3, issued 2/13/96