The OS as a time keeper

Scheduling
 In multiprogramming systems, the OS
decides which one to activate when there is
more than one runnable process (in the ready
queue).

 The decision is made by the part of the OS

called scheduler using a scheduling
algorithm.
Scheduling
 Set of policies and mechanisms to control the
order of work to be performed by a computer
system.
Scheduling
 OSmust keep the CPU busy as much as
possible by executing a (user) process until it
must wait for an event, and then switch to
another process.

 Maximum CPU Utilization obtained with

multiprogramming.

 Processes alternate between consuming

CPU cycles (CPU-burst) and performing I/O
(I/O-Burst)
Alternating Sequence of CPU And I/O
Bursts
Histogram of CPU-burst
Times
Types of Scheduler
1. Long-term (job) scheduler
– admits more jobs when the resource
utilization is low and blocks the incoming
jobs from entering ready queue when
utilization is high.

1. Medium-term scheduler (swapper)

– releases suspended processes from
memory by swapping it out when the
memory becomes over committed.
Schedulers (cont.)
3. Short-term (CPU) scheduler (dispatcher) –
controls the CPU sharing among the “ready”
processes.
Usually, the next process to execute is selected
under the following circumstances:
When a process must wait for an event
When an event occurs (I/O completion, quantum
expired)
 Scheduling Criteria
 The goal is to optimize the system performance,
and yet-provide responsive service.
 Criteria used:
1. CPU Utilization – the percentage of time that the CPU
is busy to keep the CPU as busy as possible

CPU Utilization = total process burst time x 100%

total CPU execution time
of all processes
 Scheduling Criteria
2. System throughput – number of processes completed
per time unit.

Throughtput = number of processes

total CPU execution time of all processes

3. Turnaround Time – elapsed time from the time the

process is submitted for execution to the time it is
completed.

TAT = end time – arrival time (original)

 Scheduling Criteria
4. Waiting time – amount of time a process has been waiting
in the ready queue.

WT = Start time – Arrival Time (AT)

5. Response Time – amount of time it takes from when a

process was submitted until the first response is produced,
not output (for time sharing environment). This is the time
taken by the system to respond to a user input.
Scheduling Criteria
6. Balanced Utilization – percentage of time
all resources are utilized
7. Predictability – consistency
8. Fairness – the degree to which all processes
are given equal opportunity to execute.
Avoidance of starvation
9. Priorities – give preferential treatment to
processes with higher priorities.
Note:
 Different algorithms may have different
properties and may favor one type of criterion
over the other. So, the design of a scheduler
usually involves a careful balance of all
requirements and constraints.
Optimization Criteria
 Max CPU utilization
 Max throughput
 Min turnaround time
 Min waiting time
 Min response time
Scheduling Policies
 Non-preemptive scheduling
Once the CPU is given to a process it cannot be
preempted or interrupted until its CPU Burst is
completed (no suspension)

 Preemptive scheduling
Force the current executing processes to release
the CPU on certain events such as clock interrupt,
I/O interrupts or a system call. (with suspension).
Common Scheduling
Algorithms

Non-Preemptive Preemptive

FCFS RR

Both Preemptive and Non-preemptive:

SJF and Priority
First-Come, First-Served
(FCFS) Scheduling algorithm
 Also known as First in First out (FIFO)
 Simplest scheduling policy
 Arriving jobs are inserted into the tail of the
ready queue and the process to execute next
is removed from the head of the queue.
 Jobs are scheduled in the order in which they
are received.
FCFS Scheduling algorithm
Process Burst Time
P1 24
P2 3
P3 3

 Suppose that the processes arrive in the order: P1 , P2 , P3 .

The gantt chart for the schedule is:
P1 P2 P3

0 24 27 30

 Waiting time for: P1 = 0; P2 = 24; P3 = 27

 Average waiting time: (0 + 24 + 27)/3 = 17
FCFS Scheduling (Cont.)
Suppose that the processes arrive in the order
P2 , P3 , P1
 The Gantt chart for the schedule is:

P2 P3 P1

0 3 6 30

 Waiting time for P1 = 6; P2 = 0; P3 = 3

 Average waiting time: (6 + 0 + 3)/3 = 3
 Much better than previous case
Note
 A long CPU-bound job may hog the CPU and
may force shorter jobs to wait prolonged
periods. This may lead to a lengthy queue of
ready jobs (convoy effect)
Examples:
1. Suppose that the following processes arrive for
execution, what is the CPU Utilization,
throughput and average TAT and WT for these
processes with the FCFS scheduling algorithm

a) Process BT b) Process BT
A 24 P1 4
B 10 P2 1
C 3 P3 8
Examples:
2. Calculate the TAT and WT of each process.
Consider the process arrival and burst times of
these processes as shown below.
Process BT AT Process BT AT

A 10 0 P1 8 0.0
B 5 12 P2 4 0.4
C 2 13 P3 3 8.2
D 1 18 P4 2 14.6
E 5 20 P5 1 16.0
Seatwork
 Calculate the CPU Process AT BT
Utilization, average J1 13 4
TAT and WT for the J2 5 6
execution of the J3 3 6
process below.
J4 24 10
Consider their arrival
times and their burst J5 15 7
times J6 18 7
J7 14 8
Process’ CPU and I/O
requests
 CPU and I/O requests
use any of the CPU
scheduling algorithm.

 After CPU burst, the I/O

burst follows.

 CPU burst will always be

the LAST to perform for
the completion of the
execution.
Shortest-Job-First
Scheduling
 Selects job with the shortest (expected) burst
time first. Shorter jobs are always executed
before long jobs

 One major difficulty with SJF is the need to

know or estimate the burst time of each job
Shortest-Job-First
Scheduling
 Another difficulty is long running times may
starve because the CPU has a steady supply
of short jobs

 But SJF is optimal – gives minimum WT for a

given set of processes.

 Preemptive SJF is also known as Shortest

Remaining Time First (SRTF)
Two schemes:
 Non-Preemptive – once the CPU has given to
the process it cannot be preempted until it
completes its CPU burst

 Preemptive – if a new process arrives with

CPU burst less than the remaining time of
current executing process, preempt.

Also known as the Shortest-Remaining-Time-First

(SRTF)
Example of Non-Preemptive
SJF
Process Arrival Time Burst Time
P1 0 7
P2 2 4
P3 4 1
P4 5 4
 SJF (non-preemptive)
P1 P3 P2 P4

0 3 7 8 12 16

 Average waiting time = (0 + 6 + 3 + 7)/4 = 4

Example of Preemptive SJF (SRTF)

Process Arrival Time Burst Time

P1 0 7
P2 2 4
P3 4 1
P4 5 4
 SJF (preemptive)
P1 P2 P3 P2 P4 P1

0 2 4 5 7 11 16

 Average waiting time = (9 + 1 + 0 +2)/4 = 3

Your turn
 Obtain the average waiting time and the turnaround time,
including CPU Utilization and Gantt Chart for the following set of
processes using Non-P SJF, and SRTF

Process BT AT Process BT AT
A 10 0 J1 6 11
B 6 5 J2 8 0
C 7 7 J3 12 4
D 3 6 J4 2 2
E 1 3 J5 5 5
Priority Scheduling Algorithm
(PSA)
 Priority number (integer) is associated with
each process
 CPU is allocated to the process with the
highest priority (smallest integer ≡ highest
priority)
Preemptive
Non-Preemptive
PSA
 SJF is a priority scheduling where priority
is the predicted next CPU burst time
 Problem:
 Starvation – low priority processes may never
execute
 Solution:
Aging – as time progresses increase the
priority of the process
Sample Problem #1:
Process BT PL (HP = 1)
P1 8 4
P2 1 3
P3 3 2
P4 4 1
P4 P3 P2 P1

0 2 4 5 7 8 16

Average WT: ( 8+7+4+0 ) / 4 = 4.75

Average TAT: (16+8+7+4 ) / 4 = 8.75
Sample Problem #2:
Process BT PL AT (HPL =1)
P1 8 4 0
P2 1 3 3
P3 3 2 5
P4 4 1 7

Non-Preemptive PSA: P1 P4 P3 P2

0 8 11 12 15 16

Preemptive PSA: P2 P P3 P3
P1 1 P4 P1

0 3 4 5 7 11 12 16
Round Robin (RR)
Scheduling
 Process gets a small unit of CPU time (time
quantum), usually 10-100 milliseconds. After
this time has elapsed, the process is preempted
and added to the end of the ready queue.

 If there are n processes in the ready queue and

the time quantum is q, then each process gets
1/n of the CPU time in chunks of at most q time
units at once. No process waits more than (n-
1)q time units.
Sample Problem:
Process Burst Time
P1 53
P2 17
P3 68
P4 24

P1 P2 P3 P4 P1 P3 P4 P1 P3 P3
0 20 37 57 77 97 117 121 134 154 162
Your turn:
 Given the following Job AT BT PL
HPL=1
jobs, schedule using
1 0 18 3
PPSA and RR
(QT=4). 2 8 8 2

 Answer the queries 3 10 5 2

on the next slide. 4 17 12 1
5 25 8 3
6 30 6 1
Multilevel Queue
 Ready queue is partitioned into separate
queues:
foreground (interactive) & background
(batch)
 Each queue has its own scheduling
algorithm
foreground – RR
background – FCFS
 Serves foreground first then background
Possibility of starvation.
MLQ Example
Job AT BT Type
 Assume that the CPU 1 0 5 B
scheduler uses MLQ
2 7 20 B
with 2 levels: foreground
queue for type F (jobs 3 8 1 F
uses SJF); and 4 9 12 F
background queue for 5 50 10 B
type B (jobs uses FCFS). 6 75 15 B
Scheduling algorithms
7 76 2 F
between queues is fixed
preemptive priority ( F 8 78 2 F
being the highest
priority)
Multilevel Feedback Queue
 A process can move between the various
queues; aging can be implemented this way
 Multilevel-feedback-queue scheduler defined by
the following parameters:
number of queues
scheduling algorithms for each queue
method used to determine when to upgrade a process
method used to determine when to demote a process
method used to determine which queue a process will
enter when that process needs service
Example of Multilevel Feedback
Queue
 Three queues:
Q0 – RR with time quantum 8 milliseconds
Q1 – RR time quantum 16 milliseconds
Q2 – FCFS
Multilevel Feedback Queue:
Scheduling

 New job enters queue Q0 which is served FCFS. When it

gains CPU, job receives 8 milliseconds. If it does not
finish in 8 milliseconds, job is moved to queue Q1.

 At Q1 job is again served FCFS and receives 16

additional milliseconds. If it still does not complete, it is
preempted and moved to queue Q2.
MLFQ Example
 Schedule using an Job BT AT
MLFQ:
A 10 4

Q1: RR Q=3 B 13 3

Q2: RR Q=5 C 5 0
Q3: FCFS D 6 1

E 18 1
Try it!
Job BT AT
 Schedule using an
A 24 32
MLFQ:
B 36 15
Q1: RR with Q=4
C 17 0
Q2: RR with Q=10
D 9 1
Q3: FCFS
E 15 17