An Algorithm for Optimized Cost in a Distributed Computing System

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 3004
An Algorithm for Optimized Cost in a Distributed Computing System
Safdar Alam1, Prof. Ravindar Kumar2
1 P.G Student, Dept. of Electronics & communication, Al-Falah University, Haryana, India
2 Assistant Professor, Dept. of Electronics & Communication, Al-Falah University, Haryana, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract -A distributed system consists of a collection of
autonomous computers, connected through a network which
enables computers to coordinate their activities and to share
the resources of the system. In distributed computing, a single
problem is divided into many parts, and each part is solved by
different computers. As long as the computers are networked,
they can communicate with each other to solve the problem.
DCS consists of multiple software components that are on
multiple computers, but run as a single system. The ultimate
goal of distributed computing is to maximizeperformance ina
time effective, cost-effective, and reliability effective manner.
The problem is addressed of assigning a task to a distributed
computing system. The assignment of the modules of tasks is
done statically. In this paper We study and formulate an
algorithm to solve the problem of static task assignment in
DCS,, teach task be assigned to get the more reliable results in
lesser cost. The method uses mathematicalalgorithmbyusing
optimization for optimized cost for task allocation in DCS in
MATLAB
Keywords: Distributed Network, Dynamic Allocation,
Performance, Residing cost, Reallocation cost
1.INTRODUCTION
Distributed Computing System (DCS) is a collection of
independent computers interconnected by transmission
channels that appear to the users of the system as a single
computer. Distributed systems are groups of networked
computers. The word distributed in terms such as DCS,
referred to computer networks where individual computers
were physically distributed within some geographical area.
The terms are Now days used in a much wider sense. Each
node of DCS is equipped with a processor, a local
memory,and interfaces. The purpose of the distributed
system is to coordinate the use of shared resources or
provide communication services to the users. In distributed
computing, each processor has its own private memory
(distributed memory). The processors in a typical
distributed system runconcurrentlyinparallel.Therequired
processing power for task assignment applications in a DCS
can not be achieved with a single processor.Oneapproachto
this problem is to use (DCS) that concurrently process an
application program by using multiple processors. as a
means of differentiating between the variouscomponentsof
a project. It can also be understand as usually assignedpiece
of work to the processor often to be finished withina certain
time. A task is a piece of code that is to be executed and task
allocation refers to the way that tasks are chosen, assigned,
coordinated. Execution time is the time in which single
Instruction is executed. Execution cost can be termed as the
amount of value of resource used. The execution cost of a
task depends on the processor on which it is executed
(heterogeneous processors) and the communication
between two tasks depends only on whether or not they are
assigned to the same processor (homogeneous network).
several issues arise such as the minimization of time and
cost as well as maximization of system reliability count. By
considering that the preference should be given to the idle
processor we assign load count as 1 or 0. Now, in each table
we will do the addition of each row and will also take the
average of each row on the basis of sum of each row. Now,
we will subtract the values from average. Negative and zero
values will not be considered. For time and cost minimum
value will be allocated and forreliabilitymaximumvaluewill
be considered. Now the tasks can be allocated for gettingthe
optimized results in terms of cost, also E cost can be
calculated. The function for obtainingtheoverall assignment
time, cost and reliability is as follows-Distributedcomputing
is a field of computer science that studies distributed
systems. A distributed system is a model in which
components located on networked computerscommunicate
and coordinate their actions by passing messages. The
components interact with each other in order to achieve a
common goal. Three significantcharacteristicsofdistributed
systems are: concurrency of components, lack of a global
clock, and independent failure of components. Examples of
distributed system from systems to peer-to-peer
applications.
2. Objective
The objective of this research is to find out the optimized
cost in Distributed Computing System (DCS) for a task
allocation problem or develop a task allocation model that
can minimize the overall system cost with the dynamic re-
allocation approach. task executionmustbecompletelydone
before another task takes control of the processor, and the
processor environment is homogeneous.Thismeansthatthe
processors have samespeedsor processing capabilities.This
study offers a mathematical model that allocates the tasks
dynamically as tasks executes in various phases. During the
particular task execution rest of other task are residing in
the particular phase. Execution cost for each phase [EC],
inter task communication cost [ITCC], residence cost [RC] of
each task on different processors and relocation cost [REC]

for each task are considered to design a dynamic tasks
allocation model. To achieve cost optimization in DCS
allocation method finds an allocation with minimum
allocation cost.
3. Technique
This research considers a distributed computing system
consisting of a set T = {t1, t2, t3, t4,…tm} of m tasks to be
allocated on a set P = {p1, p2, p3, …pn} of n processors
divided into k phases with criteria tasks m aremorethanthe
number of processorsn(m>n).Executioncostfor phasewise
of each processor is given in the form of Execution Cost
Matrix ECM(,,) of order k x m x n. The Residing Cost for
residing the unexecuted tasks on theprocessorismentioned
in Residing Cost Matrix RCM(,,) of order k x m x n. The Inter
Task Communication Cost between executing and non-
executing tasks are also considered and is mentioned in the
Inter Task Communication CostMatrixITCCM(,)oforder mx
k and during the processing a task is re-allocate from one
processor to another processor then it also obtained some
cost i.e. reallocation cost and it is given in the Reallocation
Cost.
Matrix RECM(,) of order m x k. To calculate ERCM(,) for each
phase sum up the values of ECM(,,) andRCM(,,).Computethe
average of each row of ERCM(,) and arrange the values in
increasing order in AVG_ROW asc (). Now it selects first n
number tasks from AVG_ROW asc() and store them in
ERCM_I(,) and remaining n number of tasks in ERCM_II by
partitioning ERCM(,) into two sub problems. Follow the
same process for next n or less than n and solve them using
assignment method. Evaluate the Execution Cost,
Communication Cost and Reallocation Cost.Followthesame
process for all phases and at the end calculate the value of
Execution Cost, Communication Cost and Reallocation Cost
to obtain the phase wise total execution cost. Calculate the
sum of optimal cost of each phase to evaluate the overall
optimal cost of distributed computing system.
4. Proposed Method
Step 1: Start Algorithm
Step 2: Input number of tasks as m
Step 3: input the number of processors in n
Step 4: Input the number of phases in k
Step 5: input the Execution Cost Matrix ECM(,,)oforderk xm
x n
Step 6: input the Residing Cost Matrix RCM(,,) of order k x
mx n
Step 7: input the Inter Task Communication Cost Matrix
ITCCM(,) of order m x k
Step 8: input the Reallocation Cost Matrix RECM(,,) of order
m x k
Step 9: For I = 1 to m
i. K = I:
ii. Add the values for ECM(,,) and RCM(,,) and store the
results in ERCM(,,)
iii. Store the average of each row of ERCM(,,) and store it in
AVG_ROW()
iv. sort AVG_ROW() in ascending and store the results in
AVG_Row asc()
v. While (All tasks of AVG_Rowasc() !=SELECTED)
{
a. Make partition of ERCM(,,) for n tasks, storeitinERCM_I(,)
and ERCM_II(,)
b. Apply assignment method on ERCM_I(,) and ERCM_II(,)
}
v. Compute Execution Cost (EC),Inter Task Communication
Cost (ITCC) and Reallocation Cost (RC)
vi. Total Cost = EC + ITCC + RC
I = I + 1
10. Applying evolutionary optimization technique
vii. Optimal Cost = (Total Cost)
Step 10: State Results
5. Implementation
This research considers a distributed computing system
which is made up of four tasks {t1, t2, t3, t4} to be allocated
on two processors {p1, p2} in five phases. The phase wise
execution cost of individual processor is known in the form
of Execution Cost Matrix ECM(,,) of order k x mxnwhere k is
the number of phases, m is the number of tasks and n is the
number of processors. Residing costs for the remaining
tasks, except for the executing task, on eachprocessorisalso
known and mentioned in Residing Cost Matrix (,,) or order k
x m x n.Inter Task Communication Cost between the
executing task and all other task if they are on different
processors also taken into the consideration and mentioned
in Inter Task Communication Cost Matrix ITCCM(,) or order
m x k. During the execution an allocated task is shifted from
one processor to another processor during the next phase
then some cost is incurred in reassignment process at the
end of each phase andit is known as reallocation cost.
Reallocation cost for the given example is also known and it
is mentioned in Reallocation Cost Matrix RECM(,)oforderm
x k.
5.1 Example
This research paper considers a distributed network system
which is made up of four tasks {t1, t2, t3, t4} to be allocated
on two processors {p1, p2} in five phases. The phase wise
efficiency of individual processor is known in the form of
Execution Cost Matrix ECM(,,) of order k x m x n where k is
the number of phases, m is the number of tasks and n is the
number of processors. The ECM(,,) is as mentioned here:
Residing costs for the remaining tasks, except for the
executing task, on each processor is also known and
mentioned in Residing Cost Matrix (,,) or order k x m x n.
RCM(,,) mentioned here:

ECM (,,) =
Phase Task Execution Cost
p1 p2
1 t1 5 4
t2 - -
t3 - -
t4 - -
2 t1 - -
t2 7 6
t3 - -
t4 - -
3 t1 - -
t2 - -
t3 3 5
t4 - -
4 t1 - -
t2 - -
t3 - -
t4 4 #
5 t1 5 6
t2 - -
t3 - -
t4 - -
Inter Task Communication Cost between the executing task
and all other task if they are on different processors also
taken into the consideration and mentioned in Inter Task
Communication Cost Matrix ITCCM(,) or order m x k. ITCCM
is mentioned in form of Table 1:
Table 1: Inter Task Communication Matrix
ITCC(,)=
Tasks ↓
Phases
1 2 3 4 5
t1 - 4 5 3 -
t2 2 - 4 5 2
t3 5 3 - 6 3
t4 2 3 4 - 0
During the execution an allocated task is shifted from one
processor to another processor during the next phase then
some cost is incurred in reassignment process at the end of
each phase and it is known as reallocation cost. Reallocation
cost for the given example is also known and it is mentioned
in Reallocation Cost Matrix RECM(,) of order m x k. RECM(,)
is shown in Table 2:
RCM (,,) =
Phase Task Residing Cost
p1 p2
1 t1 - -
t2 2 3
t3 3 2
t4 4 3
2 t1 2 3
t2 - -
t3 3 4
t4 2 5
3 t1 4 2
t2 3 4
t3 - -
t4 4 2
4 t1 2 4
t2 3 2
t3 2 3
t4 - -
5 t1 - -
t2 3 2
t3 2 3
t4 2 4
Table 2: Reallocation Cost Matrix
RECM(,)=
Tasks ↓
Phases
1 2 3 4 5
t1 2 2 5 3 -
t2 3 3 4 4 -
t3 4 3 3 3 -
t4 2 4 2 2 -
As per the execution table (ECM), task t1 will execute in
phase 1 while remaining tasks i.e. t2, t3 and t4 will be treat as
a residing tasks. On the sum up of ECM(,,) and RCM(,,) will
drive another matrix as ERCM:
ERCM (,)=
p1 p2
t1 5 4
t2 2 3
t3 3 2
t4 4 3
On evaluating the average of each row of ERCM(,) an linear
array avg_row() is obtained here:
avg_row() =
On arranging the values of avg_row() in ascending order a
new linear array avg_row_asc() linear is formed:
avg_row_asc()
=
In order to get optimal assignment, allocation techniquewill
divide ERCM(,) by selecting first two tasks from
sum_row_asc() and store the values in ERCM_I(,) and last
two tasks into ERCM_II respectively:
p1 p2
ERCM_I(,)=
t2 2 3
t3 3 2
p1 p2
ERCM_II(,)=
t4 4 3
t1 5 4
On derived two matrices i.e. ERCM_I and ERCM_II apply
assignment method to allocate the tasks and the allocation
and their costs is present in Table 3.
t1 t2 t3 t4
4.5 2.5 2.5 3.5
t2 t3 t4 t1
2.5 2.5 3.5 4.5

Ph
as
e
Ta
sk
Pr
oc
ess
or
Assig
ned
Task
Exec
ution
Cost
(EC)
Communi
cation
Cost (CC)
Realloc
ation
Cost
(RC)
Phas
e-
wise
Total
Cost
(EC +
CC +
RC)
1 t1 p1 t2 * t4 6 4 0
p2 t3 * t1 6
By applying the same process on the remaining phases
final allocation is obtained as present in Table 4 for given
example.
Table 4: Final Dynamic Allocation Table
Phase Executing
Task
Processor Assigned
Task
Phase-wise
Total Cost
(EC + CC +
RC)
1 t1 p1 t2 * t4 16
p2 t3 * t1
2 t2 p1 t1 * t4 22
p2 t3 * t2
3 t3 p1 t2 * t3 19
p2 t1 * t4
4 t4 p1 t4 * t3 24
p2 t2 * t1
5 t1 p1 t3 * t4 15
p2 t2 * t1
Total Task scheduling cost 96
6. Conclusion
This research designed a task allocation model with
dynamic reallocation technique for execution of tasks in
Distributed Computing System (DCS) and provides the
optimal solution in order to get optimized costs for task
allocation. This allocation model considered the several
factors of dynamic environment i.e. execution cost, residing
cost, reallocation cost, inter task communication cost and
most important execution phases. In dynamic model a tasks
completes its execution in various phase so presented
dynamic allocation model provide optimal solution phase
wise. The presented model is tested in MATLAB platform by
creating distributed environment as mentioned in Fig. 1.
Optimal cost is calculated for each phase and every task.
Communication cost between executing and non-executing
task, reallocation cost of task are also added to evaluatefinal
optimal cost of each phase. Phase wise results are generated
in MLATLAB for presented algorithm and algorithm [17],
results are compared for both algorithms, on comparing
phase wise execution cost and total execution cost, it is
found presented model shows the better results as
mentioned in Table 7
Table 7: Algorithm results derived in MATLAB environment
and compare with algorithm [17]
Fig.1: Distributed environment in MATLAB
Overall results also evaluated and compared with past
algorithm and found to be very less as shown in Table
Table:5 Comparison Table
7. References
[1] A.Farinelli, L. Iocchi, D. Nardi, V. A. Ziparo.2005. Task
Assignment with dynamic perception and constrainedtasks
in a Multi-Robot System, Proc. of Intern. Conf. on Robotics
and Automation (ICRA'05)
[2] Faizul Navi Khan, KapilGovil. 2014. A TRICKY TASK
SCHEDULING TECHNIQUE TO OPTIMIZE TIME COST AND
RELIABILITY IN MOBILE COMPUTING ENVIRONMENT,
International Journal of Research in Engineering and
Technology, Vol. 3 Issue 5, 823-829
[3] Faizul Navi Khan, KapilGovil. 2014. AN EFFICIENT TASK
SCHEDULING ALGORITHM TO OPTIMIZE RELIABILITY IN
MOBILE COMPUTING, International Journal of Advances in
Engineering & Technology, Vol. 7 Issue 2, 635-641
[4] Faizul Navi Khan, KapilGovil. 2014. A Static approach to
optimize time cost and reliability in Distributed Processing
Environment. International Journal of Scientific &
Engineering Research, Vol. 05, Issue 5, 1016-1021
[5] Faizul Navi Khan, KapilGovil. 2013. Cost Optimization
Technique of Task Allocation in Heterogeneous
Distributed Computing System, Int. J. Advanced Networking
and Applications, Vol. 5 Issue 3, 1913-1916
[6] Faizul Navi Khan, Kapil Govil. 2014. Cluster based
optimization routing strategy for data communication in
Total cost
Proposed
Algorithm
96
Earlier
Algorithm
115

Mobile Computing, International Journal of Computer
Applications, Volume 99, Issue 2, 19-24
[7] Faizul Navi Khan, Kapil Govil. 2013. Distributed Task
Allocation Scheme for Performance Improvement in Mobile
Computing Network, International Journal of Trends in
Computer Science, Vol. 2 Issue 3. 809-817
[8] Faizul Navi Khan, Kapil Govil, AlokAgarwal. 2014
Performance enhancementofdistributednetwork system by
Phase-wise dynamic task allocation, 2014, International
Conference on Parallel, Distributed and Grid Computing
(PGDC 2014), IEEE Proceedings, ISBN. 978-1-4799-7681-2
[9] Faizul Navi Khan, KapilGovil. 2013. Static Approach for
Efficient Task Allocation in Distributed Environment,
International Journal of ComputerApplications,Vol.81Issue
15, 19-22
[10] Harendra Kumar, M. P. Singh, P. K. Yadav.2013.Optimal
Tasks Assignment for Multiple Heterogeneous Processors
with Dynamic Re-assignment, International Journal of
Computers & Technology, Vol. 4, No. 2, 528-535
[11] Kapil Govil. 2011. A Smart Algorithm for Dynamic Task
Allocation for Distributed Processing Environment,
International Journal of Computer Applications, Vol. 28, No.
2, 13-19
[12] M.P, Singh, P.K.Yadav, H. Kumar, B.Agarwal. 2012.
Dynamic Tasks Scheduling Model for Performance
Evaluation of a Distributed Computing System through
Artificial Neural Network, Proceedings of the International
Conference on Soft Computing for Problem Solving(SocProS
2011) (Advances in Intelligent and Soft Computing:
Published by Springer ) Vol.130, 321-331
[13] Manisha Sharma, Harendra Kumar, Deepak Garg. 2012.
An Optimal Task Allocation Model through Clustering with
Inter-Processor Distances in Heterogeneous International
Journal of Computer Applications(0975 – 8887)Volume122
– No.22, July 2015 35 Distributed Computing Systems,
International Journal of Soft Computing and Engineering,
Vol. 2 No.1, 50-55
[14] Monika Choudhary, Sateesh Kumar Peddoju. 2012. A
Dynamic Optimization Algorithm for Task Scheduling in
Cloud Environment, International Journal of Engineering
Research and Applications (IJERA), Vol. 2, Issue 3, 2564-
2568
[15] N. Beaumont. 2009. Using dynamic programming to
determine an optimal strategy in a contract bridge
tournament, Journal of the Operational ResearchSociety, Vol
61, Issue 5, 732-739
[16] Martin Grajcar, “Strengths and Weaknesses of Genetic
List Scheduling for Heterogeneous
Systems”, IEEE Computer Society, Proceedingsofthe Second
International Conference on
Application of Concurrency to System Design, Page: 123,
ISBN: 0-7695-1071-X, IEEE
Computer Society Washington, DC, USA, 2001.
[17] Hesam Izakian, Ajith Abraham, Vaclav Snasel,
“Comparison of Heuristics for scheduling
Independent Tasks on Heterogeneous Distributed
Environments”, Proceedings of the 2009
International Joint Conference on Computational Sciences
and Optimization, Volume 01,
Pages: 8-12, 2009,ISBN:978-0-7695-3605-7,IEEEComputer
Society Washington, DC,
USA.
[18] Yi-Hsuan Lee and Cheng Chen, “A Modified Genetic
Algorithm for Task Scheduling in
Multiprocessor Systems”, Proc. of 6th International
Conference Systems and Applications,
pp. 382-387, 1999.
[19] Amir Masoud Rahmani and Mojtaba Rezvani, “A Novel
Genetic Algorithm for Static Task
Scheduling in Distributed Systems”, International Journal of
Computer Theory and
Engineering, Vol. 1, No. 1, April 2009, 1793-8201.
[20] Michael Rinehart, Vida Kianzad and Shuvra S.
Bhattacharyya, “A modular Genetic Algorithm
for Scheduling Task Graphs”, Technical Report UMIACS-TR-
2003-66, Institute for Advanced
Computer Studies University of Maryland at College Park,
June 2003.

An Algorithm for Optimized Cost in a Distributed Computing System

Recommended

More Related Content

What's hot (20)

Similar to An Algorithm for Optimized Cost in a Distributed Computing System (20)

More from IRJET Journal (20)

Recently uploaded (20)

An Algorithm for Optimized Cost in a Distributed Computing System