SJF Scheduling in OS

Shortest Job First (SJF) Scheduling in OS

Shortest Job First is a Preemptive or Non-Preemptive algorithm. In the shortest job first algorithm, the job having shortest or less burst time will get the CPU first. It is the best approach to minimize the waiting time. It is simple to implement in the batch operating system because in this CPU time is known in advance, but it is not used in interactive systems, because in interactive systems, CPU time is not known.

Characteristics of Shortest Job First Scheduling

1. SJF algorithm is helpful in batch operating where the waiting time for job completion is not critical.
2. SJF improves the throughput of the process by ensuring that the shorter jobs are executed first, thus the possibility of less turnaround time.
3. SJF enhances the output of the job by executing the process, which is having the shortest burst time.

Advantages of Shortest Job First (SJF) Scheduling

The advantages of Shortest Job First scheduling are:

1. SJF is basically used for Long Term Scheduling.
2. The average waiting time of Shortest Job First (SJF) is less than the FCFS (First-Come, First Serve) algorithm.
3. For a particular set of processes, SJF provides the lowest average waiting
4. In terms of the average turnaround time, it is optimal.

Disadvantages of Shortest Job First (SJF) Scheduling

1. In SJF process completion time needs to be known earlier. Although prediction is difficult.
2. Sometimes the problem of starvation occurs in SJF.
3. SJF needs the knowledge to know how long a process will run.
4. It is not easy to know the upcoming CPU request length..
5. In SJF, it is necessary to record elapsed time, resulting in more overhead the processor

Types of Shortest Job First (SJF) Scheduling

There are two types of Shortest Job First Scheduling.

1. Non-Preemptive SJF
2. Preemptive SJF

If the processor knows the Burst time of the processes in advance, the scheduling of the process can be implemented successfully. But practically it’s impossible.

When all the processes are available at the same time, then the Shortest Job Scheduling algorithm becomes optimal.

1. Non-Preemptive SJF: - In Non-Preemptive Scheduling, if a CPU is located to the process, then the process will hold the CPU until the process enters into the waiting state or terminated.

Example of Non-Preemptive SJF Scheduling:

In the following example, we have 4 processes with process Id P0, P1, P2, and P3. The arrival time and burst time of the processes are given in the following table.

Process ID	Burst Time	Arrival Time	Completion time	Waiting Time	Turnaround Time
P0	8	5	21	8	16
P1	5	0	5	0	5
P2	9	4	16	3	12
P3	2	1	7	4	6

The waiting time and turnaround time are calculated with the help of the following formula.

         Waiting Time = Turnaround time – Burst Time

          Turnaround Time = Completion time – Arrival time 

Process waiting time:

P0= 16-8=8

P1= 5-5=0

P2=12-9=3

P3=6-2=4

Average waiting time= 8+0+3+4/4

                                   =15/4

                                    =3.75  

Process turnaround time:

P0=21-5=16

P1=5-0=5

P2=16-4=12

P3=7-1=6

Average turnaround time=16+5+12+6/4

                                             =39/4

                                             =9.75

Preemptive SJF Scheduling: - In this, jobs are moved into the ready queue when they arrive. Those Processes which have less burst time begins its execution first. When the process with less burst time arrives, then the current process stops execution, and the process having less burst time is allocated with the CPU first.

Example of Preemptive SJF Scheduling: In the following example, we have 4 processes with process ID P1, P2, P3, and P4. The arrival time and burst time of the processes are given in the following table.

Process	Burst time	Arrival time	Completion time	Turnaround time	Waiting time
P1	18	0	31	31	13
P2	4	1	5	4	0
P3	7	2	14	12	5
P4	2	3	7	4	2

The waiting time and turnaround time are calculated with the help of the following formula.

         Waiting Time = Turnaround time – Burst Time

          Turnaround Time = Completion time – Arrival time 

Process waiting time:

P1=31-18=13

P2=4-4=0

P3=12-7=5

P4=4-2=2

Average waiting time= 13+0+5+2/4

                                     =20

Process Turnaround Time:

P1=31-0=31

P2=5-1=4

P3=14-2=12

P4=7-3=4

Average turnaround time= 31+4+12+4/4

                                           =12.75
The GANTT chart of preemptive shortest job first scheduling is:

Related Posts:

• FCFS (First-Come-First-Serve) Disk Scheduling Algorithm
• Evolution of Operating System
• File in Operating System
• Unix Operating System
• Components of Operating System
• Semaphore in Operating System
• File Access Methods in Operating System
• What is Demand Paging?
• Directory Structure in Operating System
• File Allocation Methods in Operating System
• Deadlock Avoidance
• Virtual Memory in Operating System