Processes in a system cycle between two states: a CPU
(Central Processing Unit) burst, in which calculations are carried out, and an
I/O (input/output) burst, in which data is either sent to or received by the system.
The Central Processing Unit (CPU) attempts to maximise the efficiency and ‘fairness’
of a CPU, by allowing a process to utilise the CPU, whilst another waits for an
I/O. If this isn’t addressed, CPU cycles are lost whilst the CPU waits for a
single process’s I/O. The role of the CPU scheduler is to find the subsequent
process for an inactive CPU, from a queue of immediately available processes.
The order in which the processes are handled depends on the algorithm implemented
by the CPU scheduler.
There are two types of algorithms applied: non-preemptive
and preemptive. Nonpreemptive algorithms must be applied applied when a new
process is required, i.e. when either a process is terminated, or when a
process transitions from a running to waiting state. However, when a process
goes from either a running state to ready state, or from waiting state to ready
state, preemptive algorithms are optimum. In order to ensure that preemptive
algorithms can be implemented, it must be ensured that the hardware can support
time interrupts. Once the scheduling algorithm has selected a process, the
dispatcher hands control of the CPU to it. This involves a context switch, a
user mode switch and then a jump to the correct location in the newly loaded
program. The time taken for this to
complete is called the dispatch latency.
When attempting to choosing the optimum algorithm to
implement in CPU scheduling, there is criteria to adhere by in order to make
the best decision: CPU utilisation should be maximised, in order not waste a
single CPU cycle wherever possible, and the ‘throughput’ (number of processes executed
per unit time) should also be maximised. However, the turnaround time, waiting
time and response time should all be minimised as much as possible. http://www.studytonight.com/operating-system/cpu-scheduling
For the following evaluation of algorithms, a single
CPU burst is assumed per process. First-Come First Served (FCFS) is the most primitive
form of scheduling, involving a first-in, first-out basis for the queue order,
and is a non-preemptive algorithm. The benefits of using FCFS is that it is
simple in application and allows the process being executed to complete its CPU
burst. Therefore, there is no need to context switch and take CPU away from
process preemptively. However, as well as having long waiting times, FCFS can also
slow the system down in a process known as the convoy effect; preventing the
CPU from being optimally used as after the large process finished, the CPU goes
through periods of high intensity processing (with a minimal I/O) and vice
Round Robin (RR) shares similarities with FCFS,
however, the CPU bursts are given limits called time quantum. If the process
completes before the time quantum limit runs out, the schedule continues in a
first in, first out manner. However, if the quantum timer limit expires, the
time interrupt occurs and the process is placed at the back of the queue (the
algorithm is preempetive). This is ideal for small tasks, as even if a large
process is placed higher in the queue than a short one, each get a ‘fair’
attempt. However, this proves problematic with processes of equal time, where
FCFS could have a lower throughput and the overhead produced through ‘context
switching’ reduces the CPU utilisation 3.
Shortest-Job-First Scheduling (SJF) arranges the
queue in ascending order of length of CPU burst time, relying on an ‘exponential
average’ for estimating each subsequent CPU burst. SJF can be behave preemtively
when a smaller process arrives than is currently being handled. The smaller process
is then prioritised and the throughput of the processes is maximised. This is regarded
as the optimum solution for CPU scheduling, however, estimating CPU bursts is
inaccurate and the algorithm is bias to smaller tasks, to the point where ‘starvation’
(low priority task is never tended to). A solution to this is to increase the
priority of older tasks in the queue (known as ‘aging’) 3.
Priority scheduling sees that processes are assigned
a value relating to the urgency of execution (can be preempetive or non-preempetive).
Values are assigned either by users or, in SJF, by finding the inverse of the
subsequent predicted CPU burst time. Priority
scheduling also suffers from problems due to starvation, but can also be
remedied using aging.
Multilevel queue scheduling can occur where
processes can be categorised and therefore assigned to a queue with a bespoke
algorithm for the task. This can be further expanded into the most flexible
type of scheduling: ‘Multilevel Feedback-Queue scheduling’, where processes may
transition from queue to queue in order to incorporate ‘aging’ and re-evaluating
process priorities. Ideal because it is suited for any scenario, however, issues
arise with complexity of the system.