CS 430 -- Operating Systems -- Dilbert Days

    General Topics

  1. (True/False) It is faster to access a random part of physical RAM than it is to seek to a random spot on the hard drive.
  2.  What is the most common reason Dilbert wants to overlap computation and I/O?

    3. Operating Systems Structures

  3. Put an 'R' next to uses of a real time operating system, an 'M' next to uses of a mulituser-at-one-time (UNIX or MVS) operating system, and an 'S' next to uses of a single user at any one time (like Windows NT) operating system. 
    ______ Control the Areane 5 rocket motor
    ______ Control the brakes on my mother's car
    ______ Run the computer that my girlfriend (she lives alone) uses to access AOL
    ______ Manage the inventory of hamburgers for McDonalds
  4. Many times real time operating systems do not have virtual memory.  Why might that be?





  5. List two functions normally found in a microkernel, and two functions normally found in a normal kernel that are typically left out of the kernel in a microkernel.
    IN _______________________________________________________________________
    IN _______________________________________________________________________
    OUT _____________________________________________________________________
    OUT _____________________________________________________________________
  6. Suppose I want to access a file on a computer managed by a microkernel.  What would I do to access the file?







  7. Suppose I wanted to write a device driver that accessed a device that could not generate a hardware interrupt.  It can be accessed by reading and writing to special parts of memory.  Could this device be used?  How might the driver work?







    8. Memory Management

  8. I want to have a multitasking operating system, but Dogbert (Dilbert's boss) is making Dilbert use a CPU without an MMU.  Dilbert is going to swap each task into RAM, run it for a time slice, and then swap the next task into the same place in RAM.  Each process has 50 pages, and dirties ten of its pages per time slice.  It takes 10 ms to read or write a page to disk.  How long does it take to swap from one task to the next? __________  (show work!!)







  9. If a time slice lasts 1 second (including transfer time), what percent of the CPU's useful work is wasted on transfering data back and forth?______





  10. (Disk/Net) Suppose Catbert has a computer that allows him to seek to any disk page in 10 ms, and transfer 1 page/ms thereafter.  He can also access a RAMdisk across a network with a 1 ms latency and 3 ms/page transfer rate.  Catbert wishes to implement a virtual memory subsystem for an application that is essentially random in its access patterns.  Which would be faster?  Why?





  11. It takes more hardware to implement pages than segments, yet most computers use pages to implement virtual memory instead of segments.  Why?





  12. (Normal/Inverted) Alice has a computer with 24 M of physical RAM and a processor with a 4G virtual address space.  Alice has ten processes each of which uses 40M of virtual address space.   A page is 1K.  A page table entry takes 24 bytes.  Linus Torvalus just had a new baby.  The month is April.  The computer's mouse has 12 buttons (ya right!).  Which will take less space to store, a direct (normal) page table or an inverted page table?

    13. Scheduling and Processes

  13. (Thread/Process) Which is quicker, to switch between two different processes or between two different threads of the same process?
  14. Circle all of the scheduling algorithms that can exhibit starvation
    Round Robin                      First Come First Serve          
    Shortest Job Next            Baby Comes Next
  15. Some schedulers divide processes into two categories.  Category one processes offer a large time quantum but are scheduled infrequently.  Category two processes have a small time slice and are scheduled frequently.  In either case the total amount of CPU cycles offered per unit time is the same.  Which category would be best for interactive tasks and why?







    16. File Systems

  16. Dilbert has a file system where each data block contains a pointer to the next block, as shown below.  His normal application will access files in a sequential order from the beginning.  Is this file system structure reasonably efficient?  Why?
    +----+       +-----+     +-----+     +-----+
    |    |------>|     |---->|     |---->|     |---->
    +----+       +-----+     +-----+     +-----+





  17. Suppose Catbert wants to access files in reverse order (last byte first, first byte last).  How does this scheme perform now, and why?





    18. Protection and Security

  18. Circle all of the following that should be protected in a multiuser system from access by normal users.
    Write access to the hard drive controller                                    Read access to the network interface
    Read access to the real time clock                                                Write access to the MMU
    The kernel's stack
  19. Suppose Dogbert and Catbert want to verify a secret password between two computers over a network that they know has snoopers.  What should they do?