See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/273132329

Challenges in Scheduling in Operating System

Conference Paper · February 2015

CITATIONS READS
0 1,297

2 authors:

Aakash Goel Gaurav Sharma

Seth Jai Parkash Mukand Lal Institute Of Engineering & Technology Seth Jai Parkash Mukand Lal Institute Of Engineering & Technology
1 PUBLICATION   0 CITATIONS    7 PUBLICATIONS   24 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Aakash Goel on 14 April 2015.

The user has requested enhancement of the downloaded file.

Bilingual International Conference on Information Technology: Yesterday, Today, and Tomorrow, 19-21 February 2015, PP. 179-182
 2015, DESIDOC

v‚ijsfVax flLVe esa le; fu/kkZj.k esa pqukSfr;ka

Scheduling Challenges in Operating System
Aakash Goel* and Gaurav Sharma
*
Seth Jai Parkash Mukund-Lal Institute of Technology, Radaur, Haryana-135 133, India
Department of Computer Engineering, JMIT Radaur, Yamunanagar, India
*
E-mail: aakashgoel12@gmail.com

lkjka”k

fu/kkZj.k v‚ijsfVax flLVe dk lcls egRoiw.kZ dk;Z gSA lHkh lalk/kuksa dk lcls vPNk mi;ksx ds fy, mi;ksx
djus ls igys fu/kkZfjr fd;k tkuk pkfg,A ,d gh dke gS rks ge vuqlwpd dh t:jr ugha gS] ysfdu okLrfod
le; ifj–”; esa ;g laHko ugha gSA rks] vuqlwpd orZeku le; esa okaNuh; gSA fdlh Hkh vuqlwpd ds eq[; dk;Z
lcls dq”ky vkSj fo”oluh; lalk/ku ds fy, jkstxkj ds volj çLrqr djus ds fy, gSA ;g Hkh jkstxkj ds volj
vkSj lalk/kuksa ds çHkkoh :i ls fu/kkZfjr ugha dj jgs gSa rks ;g dkQh iwjs v‚ijsfVax flLVe n{krk uhpk fn[kkuk
gksxk fd rF;ksa esa ls ,d gSA bl i= esa ge fofHkUu le;c)u ,YxksfjFe dk v/;;u djus vkSj vHkh Hkh le;&
fu/kkZj.k esa ekStwn gSa] tks fofHkUu pqukSfr;ksa dk irk yxkukA
Abstract
Scheduling is the most important function of operating system. All resources must be scheduled
before use for best utilization. If there is a single job then we don’t need scheduler but in real time
scenario it is not possible. So, scheduler is desirable in the present time. The main task of any scheduler
is to submit the jobs to most efficient and reliable resource. It is also one of the facts that if jobs and
resources are not scheduled effectively then it will degrade the whole operating system efficiency
drastically. In this paper are found the various scheduling algorithm have been studied and the various
challenges which still exist in scheduling.
Keywords: Scheduling, resources, jobs, scheduler

1. Introduction Scheduling of CPU resource has many ways by which

In operating system,scheduling is a basic function, it can be scheduled like FCFS, Round Robin, SJF,
since almost all computer resources are scheduled before priority scheduling. Scheduling a CPU which has
use. It involves the allocation of resources and time to different types of processes, which are required to run,
process or tasks in such a way that certain performance can be scheduled using multilevel queue and multilevel
requirements are met. In a multiprogramming, computer feedback CPU scheduling techniques. Multilevel queue
has multiple processes and they are competing for the scheduling is of much importance and is widely used
CPU at the same time. So, a choice has to be made for scheduling of jobs in a processor.
which process to run next which is done by scheduler.
Whenever the CPU becomes idle, it is the job of the
CPU scheduler to select another process from the ready
queue to run next. A scheduler executes the process with
the help of scheduling algorithms. Thus its scheduling
algorithm is heart of the OS design. A CPU scheduler
is the part of an operating system and responsible
for arbitrating access to the CPU 1. The scheduler is
mainly concerned with throughput, turnaround time,
response time, fairness, waiting time etc 2. The key to
multiprogramming is scheduling. scheduling of CPU
is one of the critical factors that affect the efficiency.
Figure 1. Scheduling behaviour.
 managing information technology

2. RELATED WORK idea of this approach is to make the operating systems

Many researchers have been proposed various adjusts the time quantum according to the burst time
methods to compare the scheduling policies in operating of the set of waiting processes in the ready queue.
system. The basic thing which is common in all of They also showed that this algorithm increases the
work related to comparison of scheduling policies is performance of RR and stability of the operating
that they have calculated average waiting time on some system and supports building of a self-adaptation
input based parameters. As waiting time is directly operating system, which means that the system is
dependent on the turnaround time and turnaround time who will adapt itself to the requirements of the user
is on execution time, so if turnaround time is high and not vice versa.
then waiting time is also high and average waiting
time also. 3. SCHEDULING ALGORITHMS
It is better to have less turnaround time. SJF tries When more than one process is ready, the operating
to have lowest execution time first. Pushpraj Singh, system must then use a CPU scheduling algorithm
Vinod Singh, Anjani Pandey 3, showed that how SJF to decide which one is to run first and for how
is better than all other scheduling algorithms. For long. Modern operating systems are moving towards
every process, average waiting time is lower for SJF multitasking environments which mainly depends on
and higher for other scheduling algorithms. They also the CPU scheduling algorithm since the CPU is the
recommended that any kind of simulation for any most effective or essential part of the computer. There
CPU scheduling algorithm has limited accuracy. The are various scheduling algorithms. A basic explanation
only way to evaluate a scheduling algorithm to code of scheduling theorems on which we had worked:
it and has to put it in the operating system, only then
a proper working capability of the algorithm can be 3.1 First-Come-First-Served (FCFS)
measured in real time systems. It is the simplest scheduling algorithm, it simply
The very advantageous thing in scheduling can be queues processes in the order that they arrive in the
if we can minimize/optimize the average waiting time. ready queue. The process which arrives first in ready
Vikas Gaba, Anshu Prashar 4, showed waiting time in queue is served first. It is a non-preemptive scheduling
genetic algorithm is less than even SJF and genetic algorithm. It means processor cannot release the process
algorithm is one of the better function optimization before its execution is over. The FCFS scheduling is
methods. This paper introduced the one new algorithm fair in the formal sense or human sense of fairness
i.e., genetic algorithm and making a comparison of but it is unfair in the sense that long jobs make short
it with SJF and FCFS scheduling algorithms based jobs wait and unimportant jobs make important jobs
on the average waiting time of these algorithms at wait 7,8.
different number of iterations.
It can also be possible that a process may demand 3.2 Shortest Job First (SJF)
for more I/O than the CPU. So, we need an efficient The algorithm gives priority to shortest process
scheduling to compensate CPU process with I/O available in ready queue.It works under the two schemes
process. Fo r this, Sanjaya Kumar Panda, Debasis (preemptive and non-preemptive). The main problem
Dash, Jitendra Kumar Rout 5, proposed group-based with this discipline is the necessity of the previous
time quanta (GBTQ) algorithm, which shows less knowledge about the time required for a process to
waiting time than SJF. A GBTQ is proposed in this complete. The process which has high burst Time
algorithm. Each group has different time quanta. (BT) has to suffer most in this algorithm. This type
This algorithm can extend to I/ O processes. It also of suffering is called as ‘aging’.
reduces starvation as well as context switching (CS).
Selection of time quanta in Round Robin (RR) plays 3.3 Round Robin (RR)
a very big role in deciding average waiting time. Scheduling is the based on time sharing concept.
If time quanta is extremely large, the RR policy is It is basically the combination of first come first
same as FCFS policy. If time quanta is extremely served (FCFS) scheduling algorithm and preemption
small, the RR approach is called processor sharing among processes. To implement the RR scheduling, we
and creates the appearance that each of n process has keep the ready queue as FIFO queue of processes.
its own processor running at 1/n the speed of real As new process arrives, it will be inserted in to
processor 8. So, the selection of time quanta can be the tail of the ready queue.
made easy by choosing dynamic time quantum. For
this, Abbas Noon, Ali Kalakech, Seifedine Kadry 6 3.3 Min-Min
proposed propose a new algorithm, called AN, based Min-min algorithm finds the task which has
on a new approach called dynamic-time-quantum; the minimum execution time and assigns the task to the

180
 Goel & Sharma : Scheduling Challenges in Operating System

resource that produces minimum completion time. The scheduler selects the job from ready queue & finally
Min–min scheduling algorithm chooses the smaller jo b is given to CPU. But the current challenge is
task first. Min–min scheduling prefers the minimum that before entertaining to any job by scheduler,
completion time. Main drawback of this algorithm is it doesn’t look for the type of job it is. Basically,
unbalanced load. there are two types of job. One is computational
intensive and other is communication intensive. If
3.4 Max-Min the job is communication intensive then it would
The Max-min scheduling algorithm is similar to
require higher bandwidth for the transfer of data
Min-min scheduling algorithm but it schedules the
larger task first. Min-min and Max-min are used for
and if it is computation intensive then it would
small scale distributed system. Thus, Max-Min would require higher CPU speed. The computational jobs
try to schedule at the same time all the short jobs and are the jobs that require processing time more than
the longest ones while Min-Min would schedule first communication time. However, communicative jobs
the shortest jobs and after that the longest ones. are the jobs that require communication time more
than processing time. In computational intensive,
3.5 Min-Max job considers execution time for processing and
Min-max heuristic has two steps for assigning it also takes more power.
jobs to machines. In first step, similar to min-min
method, completion time of all unassigned tasks (1 4.2 Reliability
≤ j ≤ n) on all the available machines (1 ≤ i ≤ m) The primary concern in systems is of security
is used to calculate the minimum completion time and resources. A proper safety mechanism is
(MCTi) of task Ti on machine Mi*. In the second
required for grid resources, because if resource is
step for all tasks ratio of minimum execution time
affected by virus or some malicious code then it
(time to execute on fastest machine) to its execution
time on selected machine Mi* is computed and the
will not execute the application safely and degrades
task which has maximum value pf it is selected for the performance of the system 9. The reliability
assignment. factor (RF) value of each system is calculated
through its self-protection capability and reputation
3.6 Comparisons of Algorithms weightage obtained from user community on its
FCFS scheduling algorithm is very easy to understand past behaviour.
and it is particularly troublesome for time-sharing
systems, where each user get a share of CPU at 4.3 Location Dependency
regular intervals. So, it is very unfavorable to allow In our distributed operating system, communication
one process to keep the CPU for long period. The jobs should be submitted at nearby location. By this,
SJF scheduling algorithm is to serve all types of job there would be no network congestion problem and
with optimum scheduling criteria. The process with packet delay has dropped to considerable rate which
less execution time is served first and the process with results in low traffic. If communication jobs are to
highest execution time is not served at all which gives be submitted at very far location from system then
minimum average waiting time and average turnaround it would make network congestion problem. In this,
time. But, it cannot be implemented at the level of network problem increases and also increase the
short-term CPU scheduling. The Round Robin (RR) failure rates.
scheduling algorithm found as most widely adopted
algorithm which may sometimes go under severe 5. CONCLUSIONs
problems directly related to quantum size. The main In this paper, we have studied the Scheduling
point is that RR is more responsive than shortest job algorithms & find the challenges encountered during
first (SJF). So, for large number of jobs, SJF gives the scheduling of jobs. Scheduling is a major area
better average waiting time than other. For small no in the operating system. We determined that during
of jobs, there is race between FCFS & SJF but still scheduling, operating system doesn’t consider the types
SJF is better. of job it is. It simply applies the algorithm on the
basis of execution time, no of processes, etc. We can
4 CHALLENGES IN SCHEDULING minimize the average waiting time more effectively if
4.1 Type of Job we are known of the type of job, i.e.,communication
When the process is allocated to CPU, Scheduling or co mputational. Location dependency is also one of
of processes starts. At present, long term scheduler the major factors deciding the efficiency of scheduling
simply takes the job in ready queue & short term algorithms. The effective and efficient exploitation

181
 managing information technology

of grid computing facilities needs highly advanced optimum multilevel CPU scheduling algorithm.
and protected resource management systems. Efficient IEEE, 2010. 978-0-7695- 058-0/10.
resource sharing and accessi ng cannot go without the 2. h t t p : / / e n . w i k i p e d i a . o r g / w i k i / S c h e d u l i n g _
assurance of high trustworthiness. (computing).
3. S i n g h , P u s h p r a j ; S i n g h , Vi n o d ; a n d
fu"d"kZ Pandey, Anjani. Analysis and comparison of CPU
scheduling algorithms. IJETAE, 4(1), 2014
bl i= esa] fu/kkZj.k ,Yxksfjne dk v/;;u fd;k vkSj
4. Gaba,Vikas, and Prashar,Anshu, Comparison of
ukSdfj;ksa ds fy, le;&fu/kkZj.k ds nkSjku lkeus pqukSfr;ksa
processor scheduling algorithms using genetic
dk fey pqds gSaA fu/kkZj.k v‚ijsfVax flLVe esa ,d çeq[k
approach ijarcsse, 2(8), 2012.
{ks= gSA ge fu/kkZj.k ds nkSjku] v‚ijsfVax flLVe ;g gS dke 5. Panda, Sanjaya Kumar.; Dash, Debasis.; and
ds çdkj ij fopkj ugha djrk fd fu/kkZfjrA ge ukSdjh Rout, Jitendra Kumar, A group based time quantum
;kuh lapkj ;k dEI;wVs”kuy ds çdkj ds fy, tkus tkrs round robin algorithm using Min-Max spread
gSa] vxj ;g cl fu’iknu le; ds vk/kkj ij ,YxksfjFe ij measure. Inter. J. of Comp. App. 64(10), 2013.
ykxw gksrk gS] çfØ;kvksa vkfn dk dksbZ ge vkSj vf/kd çHkkoh 6. Noon, Abbas.; Kalakech, Ali.; & Kadry, Seifedine
<ax ls vkSlr çrh{kk ds le; dks de dj ldrs gSaA LFkku A new round robin based scheduling algorithm
fuHkZjrk Hkh le;&fu/kkZj.k ,Yxksfjne dh n{krk esa fu.kZ; ysus ds for operating systems: Dynamic quantum using
çeq[k dkjdksa esa ls ,d gSA fxzM daI;wfVax lqfo/kkvksa dk çHkkoh the mean average. IJCSI Inter. J. Comp. Sci., 8(3),
vkSj dq”ky “kks’k.k vR;f/kd mUur vkSj lajf{kr lalk/ku çca/ku 2011.
ç.kkyh dh t:jr gSA dq”ky lalk/ku lk>k djus vkSj igq¡pus 7. Silberschatz ,Galvin and Gagne. Operating systems
mPp fo”oluh;rk ds vk”oklu ds fcuk ugha tk ldrsAA concepts, 8 th edition, Wiley, 2009.
8. Bawa ,Rajesh Kumar, and Sharma , Gaurav,.
References Reliable resource selection in grid environment.
1. Sindhu, M.; Rajkamal, R. & Vigneshwaran ,P.An Inter. J. Grid Com. App. IJGCA, 2012, 3(1).

182

View publication stats