The primary goal of a modern operating system is (consider current perspective and not a historical perspective)
to make the underline hardware (the computer system) more convenient to use
Which of the following is typically the critical performance requirements in an interactive (time-sharing) operating system enviorment?
minimizing response time
Consider the multi-level feedback unique processor scheduling protocal. Which type of process is favored by this protocal?
I/O - bound processes
A standard "garden-variety" context-switch involves removing a running process from the CPU and replacing it with another process. Assuming that a dispatcher process is responsible for handling the context-switch and no interupt handler is required for this context-switch, how many different "contexts" will the CPU see before the context-switch is completed?
In problem #4, we assumed that no interupt handler was involved in the context-switch. In this case, assume that an interupt handler is involved with the context-switch (assume, for example, that the executing process requested some I/O service). Assuming that a dispatcher process is responsible for handling the context-switch, how many different "contexts" will the CPU see before context-switch is completed?
The long-term scheduler is responsible for placing processes into the job pool (or process pool) based upon some selection protocoal
The short-term sceduler is responsible for removig processes from the ready list and placeing them in the correct I/O device waiting queue
Which of the following items least likely to be included in the PCB for a process
the numberr of interrupts generated by the process thus far
Historically, the design of operating systems has focused on:
efficient use of system resources
An I/O bound process is
a process that requires more thime in I/O service than it does in CPU service
Which of the following is not a class of interrupts?
Although what the CPU does in response to either an interrupt or a trap may be very similar, a trap is fundamentally differnt from an interrupt in that:
a trap relates to an error or exception condition generated within the currently running process
Dual-mode operation allows the OS to pretect itself and other system components. the two modes associated with dual mode operation are:
user mode and kernel mode
The cache memory (if there is any at all) is invisible to the operating system.
One of the main advantges to the layered approach to the OS design is:
the system can be constructed in a modular fashion
Which of the following is not a goal of uniprocessor scheduling?
increase memory utilization
Consider a system in which threads are supproted only at the user level (the kernel is not threaded and supports only processes). Suppose we have twow processes, P1 consissting of a single thread and P2 consisting of 50 threads. Assuming that each process receives the same length time slice in the CPU, how much slower is a thread in P2 executed thatn the tread in P1
50 times slower
Which of the following not a criterion commonly associated with uniprocessor scheculing protocols:
not cpu utilization
not response time
not waiting time
General priority scheduling of the CPU suffers from which of the follwing problems
The SPN (Shortest Process Next) scheduling protocol prioitizes processes based upon the length of their expected service time. Since this basically means that you must predict the future, such a protocol cannot actually be implemented and is useful only as a theoretical tool.
General priority scheduling can lead to starvation (infinite blocking). To alleviate this problem, many different sollutions have been proposed including:
aging the process
Processor scheduling techniques that use a feedback mechanism are basically penalizing processes that have been executing for long periods of time, thus tending to prefer processes which have not been executing for long periods of time.