ISSN: 2455-2631 © May 2016 IJSDR | Volume 1, Issue 5
IJSDR1605010 International Journal of Scientific Development and Research (IJSDR) www.ijsdr.org 49
Comparison between EDF_RM and EDF_DM in
dynamic scheduling algorithm with sporadic task
Hemang Thakar
Assistant Professor
DJMIT engineering college
Computer Science and Engineering
Abstract: Real time system is divided in to two
algorithms static and dynamic. Both algorithm want to
complete his job in given deadline or before deadline but
no one algorithm are successfully complete his job in
overloaded and under loaded condition. Dynamic
algorithm able to give optimal result in under-loaded
condition but not able to give same result in over-loaded
condition. Static algorithm not able to give optimal
result in under-loaded condition but able to perform
well in over-loaded condition. So we think new algorithm
to give result optimum in under-loaded condition and
give high result in over-loaded condition. We are
thinking to mix up dynamic scheduling algorithm EDF
(Earlier dead line first) and static scheduling algorithm
DM (dead line monotonic). EDF algorithm give optimal
result in under-loaded condition and DM algorithm
perform well in over-loaded condition. In this paper we
compare EDF_RM and EDF_DM algorithm. Here
EDF_RM is better than EDF_DM.
Keywords: EDF, LLF, RM, DM, Scheduling Algorithms,
Real-Time operating Systems.
I. Introduction
Real-time systems have well defined, fixed time constraints.
Two types of real-time system available hard and soft.
Where hard real time system work on principal job must
complete their execution before deadline. It’s useful where
missing deadline may be serious consideration. Soft real
time system is more flexible which allow job to execution
beyond the deadline. It’s applicable where cost of job is
more important than can sequences of missing deadline.
a. Real-Time Scheduler
Generally it’s worked in two categories static and dynamic.
In case of static algorithm its applied priorities at initial
(design) level and remain same throughout the task. In case
of dynamic algorithm priority assign on runtime which
depended on parameters of task. We can make dynamic
scheduling with static priority where Rate Monotonic (RM)
and Deadline Monotonic (DM) are two examples of its and
Earliest Deadline First (EDF) and Least laxity (Slack) time
First (LLF) are belongs to dynamic scheduling with
dynamic priority. Here in dynamic priority algorithm are
implements in categories first. In which job to change
priority daring it is activation and in another type job cannot
able to change priority. EDF is job level fixed priority
algorithm of this category. Another side job can change
priority during execution called job level dynamic-priority
algorithm. Least laxity (slack) time algorithm belongs to it.
At time t, the slack time of a job is (d - t), where d is the
jobs deadline and t is its remaining execution requirement.
Here, the slack time is the maximum amount of time a job
may be forced to wait if it were to execute on a processor
and still meet its deadline. The LSF algorithms give higher
priority to jobs with smaller slack time. Since the slack time
of a job can change over time, the job priorities can change
dynamically.
II. The Scheduling Algorithm
a. EDF Scheduling Algorithm
EDF Scheduling Algorithm is called earliest dead line first
or nearest deadline first Scheduling Algorithm. EDF is
dynamic Scheduling algorithm. The task with the earliest
deadline has the highest priority [1]. EDF Scheduling
Algorithm gives 100% task utilization when system is in
under loaded situation or U<=1 But when task utilization
cross load factor more then 1 or slightly overloaded
processor utilization decreases exponentially [4].
b. RM Scheduling Algorithm
RM Scheduling Algorithm is called Rate Monotonic
Algorithm.RM is a fixed or static priority scheduling
algorithm. RM assigns priorities to tasks based on their
periods [4]. Disadvantage of this algorithm is that they are
not give 100% result in under loaded situation [6]. RM gives
better performance in over loaded situation as compare to
dynamic scheduling. In RM algorithm shortest period gives
first chances to execute but when more than one task have
same period then RM randomly selects one for execution
next [7].
c. DM Scheduling Algorithm
DM Scheduling Algorithm is called Deadline Monotonic.
DM also fixed or static priority Scheduling Algorithm. DM
Scheduling Algorithm is depends on priority. In DM
algorithm shortest priority gives first chances to execute but
when more than one task have same priority then DM
randomly selects one for execution next [2]. In DM
Algorithm deadline is greater than execution time. DM is
not give 100% result in under loaded situation. But DM
gives better performance in overloaded situation as compare
dynamic scheduling.
d. EDF_RM Scheduling Algorithm
EDF scheduling algorithm perform optimal in under loaded
situation and RM scheduling algorithm perform better