Proceedings of the Third International Conference on Trends in Electronics and Informatics (ICOEI 2019)

IEEE Xplore Part Number: CFP19J32-ART; ISBN: 978-1-5386-9439-8

A Survey on Task Scheduling using Intelligent

Water Drops Algorithm in Cloud Computing
Vinay Ahari Venkatesan R
Department of ComputerScience, Department of ComputerScience,
Karunya Institute of Technology and Sciences, Karunya Institute of Technology and Sciences,
Coimbatore, India Coimbatore, India
[email protected] rlvenkei2000@g mail.co m

D Ponmary pushpa latha

Department of Information Technology,
Karunya Institute of Technology and Sciences,
Coimbatore, India
[email protected] m

Abstract— In recent years, Cloud computing emerges as an applications and remote date accessing procedure, the need for strong
innovative platform that provides a high-performance computing task scheduling algorithms becomes more important in data centers.
environment to the users and also provides the delivery of on- Scheduling in cloud computing is categorized into different classes
demand computing resources. Efficiency of cloud system is i.e., based on Task and on various metrices used in cloud. Based on
enhanced by using various resource management methodologies. task scheduling, Various classification of scheduling algorithm has
Despite the hype, S cheduling is still remaining as a challenging been identified like static scheduling and dynamic scheduling. Task
task. The main aim of this research paper is to discuss about the are arrived concurrently to the processor and also submitted to
comparison of IWDs based algorithm with many familiar arrives simultaneously at the processor in static scheduling. The
scheduling algorithms, such as MIN-MIN, MAX-MIN algorithm, tasks are submitted to the obtainable resources and the scheduling
Round Robin algorithm, FCFS algorithm, MCT algorithm, PS O decisions are made ahead of the tasks are submitted. The processing
algorithm and C-PS O algorithm where the IWD algorithms time of the task is noted after the completion of the task [5].
presents enhancement in the functional output and economy in Following parameters are not fixed in dynamic scheduling: resource
many circumstances. location, machine location and number of tasks. Static scheduling
does not aware of the task’s arrival time before submission. FCFS,
Keywords— Cloud computing, Intelligent water drops, round robin, min-min, max-min. etc. can be implemented through
Optimization, S warm intelligence, Task scheduling static scheduling algorithm. All heuristic scheduling algorithms and
methods like, Genetic algorithm, Particle Swarm Optimization
technique, Simulated Annealing algorithm, IWD algorithm, Ant
I. INT RODUCT ION
Colony Optimization technique and dynamic list scheduling
Cloud computing is a widely used storage technique to avail the algorithm, all these are coming under dynamic algorithms. Following
services and data remotely to the users. It also provides the are the metrics which are used to calculate the Quality of Service
accessibility from anywhere and at any time. Due to its flexible and (QoS) in the cloud: 1. M ake span, 2. total execution time, 3. early
dynamic infrastructure, many administrations and establishments are completion time, 4. cost, 5. fairness, 6. trust, 7. deadline, 8. response
transferring their data values and required applications to the cloud time, 9. reliability and 10. availability [6].
[1]. It also provides configurable, rapidly provisioned software This research paper is designed to analyze the various types
services with QoS guaranteed computing environment. Hence, large of task scheduling algorithm which are used by cloud environment
number of operational research areas in cloud computing are storage, Figure 1. Various existing task scheduling algorithms for computing
security, performance, load balancing, resource utilization, task environments are facing difficulties when it is applied in the cloud
scheduling, privacy, workflow scheduling, energy management, etc. environment since the resource pools are constrained by the cost plan
[2] [3]. offered by the cloud provider. The purpose of various task
M apping the tasks with available resources in the cloud is the scheduling algorithms is to reduce the processing cost along with
major process of scheduling. At the same time the mapping will be optimized resource utilization [7].
carried out according to requirements and characteristics of tasks.
Efficient Task scheduling is one of the most sensitive issue and it
influences different characteristics of cloud computing environment
such as performance, resource utilization, cost, make span time, etc.
[4]. When there is a hike in the number of users to use complex cloud

978-1-5386-9439-8/19/$31.00 ©2019 IEEE 39

 Proceedings of the Third International Conference on Trends in Electronics and Informatics (ICOEI 2019)
IEEE Xplore Part Number: CFP19J32-ART; ISBN: 978-1-5386-9439-8

necessary to prioritize schedule the tasks in cloud. The main

objective is to implement the schedule mechanism with the
tasks in cloud. This schedule mechanism was studied by
several authors who have proposed several algorithms to
schedule tasks over cloud environment.
Authors in [16] have used Ant Colony Optimization
(ACO), which is designed for achieving the user-specified
QOS constraints which are having reliability constraints,
Makespan constraints, cost constraints, reliability
Fig. 1: Scheduling in Cloud Computing optimization, Makespan optimization, cost optimization.
Artificial ants are selecting the heuristics based on pheromone
II. OVERVIEW OF INT ELLIGENT WAT ER DROPS(IWD) values in order to use the adaptive scheme.
Travelling salesman problem [8], Graph Coloring [9], Author name Pandey S et al., [17] have used Particle
Vehicle routing problem [10], Steiner tree problem [11], Swarm Optimization (PSO), this algorithm consider the cost
Multidimensional Knapsack problem [12], Economic Load of scheduling based on computational and data transmission in
Dispatch [13], Air Robot Path Planning [14] etc., are the the cloud resources.
various problems where optimization has the major role and
that has been rectified by IWD. IWD algorithm has been Heterogeneous and dynamic grid environment task
devised by Shah-Hosseini, H. in 2007 [8].IWD algorithm is scheduling are provided by the Immune Particle Swarm
designed in such a way that, water stream has chosen the Optimization (IPSO) [18]. It addresses multi-objective
optimal path to touch its destination which may be a lake or optimization problems in grid environments using objective
ocean. Water drops in lakes, rivers, and seas are the idea function based on satisfaction rate. The researchers [19]
behind the IWD algorithm. During the movement to the proposes an market-oriented hierarchical scheduling which
target, the water drops respond to the environments such as considers Task-to-Service assignment, and the Task-to-VM
river beds. Water drops can change the river beds and also assignment and optimize in local cloud data centers.
influence their direction. When there is no disturbance, the Preliminary experiments using SHEFT scheduling algorithm
water drop reaches the destination. There are various types of are discussed by the authors in [20], It does not only show that
barriers are affecting the flow and they are numbered as it outperforms different scheduling algorithms in optimization
follows: 1. rocks, 2. twists and 3. turns. of execution time, It also provides the platform to scale the
resources elastically during execution.
The Natural water drops of IWD algorithm have affected
by two major propertie and that are as follows: 1. the velocity Researchers [21] designed a Mulitple QoS constrained
of the water drops, 2. the amount of soil in their path. The Scheduling Strategy which has the capacity that can able to
movement of water drops along the path is affected by the schedule various tasks at any time. It satisfies the requirements
of QoS rwhich are needed to schedule multiple workflows.
amount of soil, they tend to move quicker along the path with
Another algorithm based on PSO was presented by authors in
a smaller amount of soil. At increased speed water drops can
[22] shows that, schedule applications among cloud services
erode more soil from the path. In choosing the path, the consider data transmission and computational cost,
amount of soil has the major problem, since the water drop Performance of makespan and cost optimization are increased
will select the effortless path largely ie the path which are in this algorithm. The Cat Swarm Optimization (CSO) in [23]
having small amount of soil. aimed at reducing wastage of energy. This algorithm receives
There are three significant reasons for IWD algorithm: (1) the optimum solution in small amount of iterations and the
efficiency of this algorithm is improved rather than PSO.
IWD solutions are quicker than the other techniques (2) High
quality solutions are obtained using average values, (3) This Following are the various task scheduling algorithm
algorithm works well in the dynamic environment and which are listed as follows: FCFS [24], Round Robin [25],
includes pop-up threats in more easiest way [15]. Min- Min [26], Max-Min [27], Chicken Swarm Optimization
[28], Improved Raven Roosting [29], Artificial Bee Colony
III. RELAT ED W ORKS [30], Cuckoo Search Algorithm [31], Bat Algorithm [32] etc.,
Cloud computing also called as heterogeneous system,
because it stores large volume of data. Now a days Cloud FCFS are working in the principal of first come first serve
Computing is an up-and-coming technology. So, it is used to basis in VM. The algorithm was simple, and execution was
advance the usage of resource in cloud, decreasing the found to be faster. Here the arrival time is considered for task
processing cost, improve the performance of the server, scheduling and less utilization of Virtual Machines are not
minimizing the processing and completion time. it is very considered.

978-1-5386-9439-8/19/$31.00 ©2019 IEEE 40

 Proceedings of the Third International Conference on Trends in Electronics and Informatics (ICOEI 2019)
IEEE Xplore Part Number: CFP19J32-ART; ISBN: 978-1-5386-9439-8

Round Robin is the pre-emptive process scheduling Bat Algorithm was projected for continuous optimization
algorithm which provides small unit of time for execution of problems. The algorithm works under the concept of
every tasks with equal probability. The response time, arrival echolocation behaviour of bats. It has number of convergence
time, load balance increases the efficiency of scheduling. properties on diverse standard functions and deals with
Tasks which are having small amount of execution time is optimization problems.
chosen for all available tasks to execute in Min-Min Obtaining the Quality of Service in the elastic and
algorithms. It has better makespan but has Poor QoS and load heterogeneous cloud environment are the area where more
imbalance. Cuckoo search is an algorithm which works well in contribution is needed even though, the number of researchers
optimization and the local search also done more efficiently contributed much in the scheduling.
using this algorithm. In Cuckoo Search, the solutions are
rationalized in the search space towards continuous valued
IV. COMPARAT IVE STUDY OF TASK SCHEDULING A LGORIT HMS
positions and the maximum time has been taken by this
algorithm to have the optimal solution. After the text edit has been completed, the paper is ready
Table 1. provides a summary of few task scheduling
Improved Raven Roosting is an optimization algorithm
algorithms in cloud computing environment. This survey
which is motivated by community behaviour of raven and the provides the areas where the existing algorithms are needed to
information is flowing from one member to another member be improved with better performance on Cloud scheduling.
in order to find the food. Premature convergence is solved by The scheduling algorithms behave differently for private,
this algorithm and it has improved performance when the public, community and hybrid cloud when it uses various
search space is high. deployment models. The parameters like bandwidth,
Chicken swarm optimization algorithms works well by communication cost etc. may vary from one deployment
model to others.
finding the food in a group of chickens using hierarchical
behaviour. Balance among the local and global search is As the various research shows that, PSO cannot perform
maintained by using the optimum solution. when search space is large and it falls into global optima. In

Algorithm Description Parameters Advantages Disadvantages Future Work

Intelligent Finite time steps are Resource 1.The speed of this The power of the The energy usage
Water Drops mainted to move the utilisation size, algorithm is faster proposed algorithm of the resource
Algorithm water drops from source cost, time, comparatively has been reduced and optimization
(IWD) to destination. Water Number of data- with other since will be
drops always prefer the centers, number algorithm. homogeneousVM s concentrated more
passageway with lesser of Virtual 2. The algorithm creates no as a future work.
soil. IWD algorithm M achine’s, PEs provides high preference of any
develops probabilistic number (CPU), quality solutions VM over the others.
solution for the optimum Bandwidth, RAM by means of
path(s). per VM average values.
3. This algorithm
works well in the
dynamic
environment and
also includes the
popup threats
without difficulty.
4.IWDC provides
reasonable costs
and stable
behaviour.

FCFS (First The tasks which comes Arrival time The algorithm is 1.The algorithm is Future work is to
Come First first in a FIFO queue simple and also using non-pre- reduce execution
Serve) will be executed first on gives fast emptive scheduling. time and reduce
VM . execution. 2. M ain component less utilisation of
of this scheduling is resources
arrival time, and if
some VM have a
smaller amount
utilisation it doesn’t
consider that.
3.The shortest task
which are at the
back of the queue
will have to wait
for the long task at
the front to finish.
 Proceedings of the Third International Conference on Trends in Electronics and Informatics (ICOEI 2019)
IEEE Xplore Part Number: CFP19J32-ART; ISBN: 978-1-5386-9439-8

order to increase the performance of PSO and IWD, the

authors propose implementation of Hybrid PSO-IWD
algorithm for task scheduling and expect that it will reduce the
cost, makespan time, execution time, response time, increase
throughput etc. and will provide better results with some
additional parameters such as to increase energy efficiency.

978-1-5386-9439-8/19/$31.00 ©2019 IEEE 42

 Proceedings of the Third International Conference on Trends in Electronics and Informatics (ICOEI 2019)
IEEE Xplore Part Number: CFP19J32-ART; ISBN: 978-1-5386-9439-8

RR (Round The jobs which are Response time Since the jobs will If the load of the Increase response
Robin) available in queue ring have their own resource is heavy time and reduce
are having the same turn, there is no then it takes much time to execute
execution time and they need for the jobs longer time to the tasks
are executed in their to wait for their execute all the jobs
turns. The job which are previous jobs in queue.
available in ring is not completion time.
executed during its turn
then it waits for its turn
to get executed.

M AX-M IN The task which is having M inimum Unfair resource It Reduces It Improves
maximum execution makespan, load allocation is done makespan and load makespan and
time is executed first. balancing, among tasks. balance across load balancing
completion-time, resources. among resources
service-time.

M IN-M IN M inimum execution M akespan The algorithm 1. This algorithm It improves load
time job is given for creates more has the problem in balance and
execution. efficient span. Load Imbalance and improvement on
having trouble in QoS parameters
maintaining QoS .
2. Smaller
makespan as tasks
are scheduled on the
fastest machines
where they are
completed earlier.

Particle Optimum minimum Inertia, Total 1.High utilization The algorithm Other parameters
Swarm values are found by this execution cost of resources. speed is slow when need to be
Optimization algorithm to schedule and thetotal 2.This algorithm the search space is considered such as
the tasks in order to executiontime, has improved high. storage cost and
avail the correct makespan computational transfer time
resource. time rather than
other scheduling
algorithm GA
(Genetic
Algorithm) and
ACO (Ant Colony
Optimization)..

Ant Colony General concept about QoSparameters, 1.It is Random It is slower as There is a need to
Optimization The idea behing the Ant makespan Guided search compared to IWDS work to reduce
Colony Optimization constraints, algorithm which and some other more cost and
algorithm is to imitate Reliability satisfies user algorithms makespan
the behaviour of ants. constraints. deadline constraints and
When it searches the constraints and speedup.
food source, it minimize the cost
communicates with of execution of
other using a chemical Tasks.
that is called as 2.M akespan
pheromone. optimization,
Energy efficiency,
server utilization
etc.

Artificial Bee It was stimulated via Food sources, It is Strong It does not perform It improves global
colony the intelligent foraging employed robustness, fast well in terms of search ability,
optimization behaviour of foragers, and convergence and makespan. Enhancement of
algorithm Honeybees. In ABC, unemployed high flexibility robustness
artificial forager bees foragers and having fewer
(agents) search for rich control parameters
artificial food sources
Artificial bees move
towards better solutions
by means of neighbour
search mechanism while
abandoning poor
solutions. 978-1-5386-9439-8/19/$31.00 ©2019 IEEE 43
 Proceedings of the Third International Conference on Trends in Electronics and Informatics (ICOEI 2019)
IEEE Xplore Part Number: CFP19J32-ART; ISBN: 978-1-5386-9439-8

enhancements in algorithms to work well on optimizing the

energy usage of various types of cloud resources. The authors
V. PROPOSED SYST EM conclude that, for homogenous Virtual Machines, there is a
The proposed research work has been displayed in a need for the improvement of IWDS algorithm. Also, there is a
need for improving various scheduling algorithms in this field.
framework in Figure 2. The tasks which are arriving at the
Hence there is a great need for task scheduling algorithms for
data centers of the cloud for execution are having the mapping the tasks in order to minimize the cost, and makes
dissimilar types and sizes. In this research work, the task has improvement in several other parameters.
been given as the input, and it is managed and directed to
arrive into the cloud data centers. This research work uses
Particle Swarm Optimization Algorithm (PSO) to quickly
REFERENCE
distribute tasks to Task scheduler efficiently. These tasks are
[1] S. Sadhasivam, “Improved Cost -Based Algorithm For Task Scheduling
gathered and formed into task queues. In,” 2010.
[2] A. V Karthick, E. Ramaraj, and R. G. Subramanian, “ An Efficient
Multi Queue Job Scheduling for Cloud Computing,” 2014.
Cloud data centers have variety of resources such as CPU, [3] Pop F et al. Reputation guided genetic scheduling algorithm for
bandwidth and memory and those algorithms are addressed by independent tasks in inter-clouds environments. In: 2013 27th
international conference on advanced information networking and
the proposed algorithm using Hybrid PSO-IWDs algorithm. In applications workshops (WAINA). IEEE; 2013.
this algorithm, the task has been taken from the task queue in [4] Alexander Visheratin, Mikhail Melnik, Nikolay Butakov and Denis
order to execute it. After that, it checks the status of Virtual Nasonov “Hard-deadline constrained workflows scheduling using
metaheuristic algorithms” YSC 2015. 4th International Young
Machines, if there are homogeneous virtual machines present, Scientists Conference on Computational Science ITMO University, St.
then the PSO algorithm is used to allocate resources and Petersburg, Russia. Published by Elsevier B.V. © 2015.
[5] Salot, Pinal. “A survey of various scheduling algorithm in cloud
execute tasks, as IWDs has limitations with homogeneous computing environment.” International Journal of Research in
Virtual Machines. For, heterogeneous virtual machines, the Engineering and T echnology 2.2 (2013): 131 -135
[6] Shaymaa Elsherbiny, Eman Eldaydamony, Mohammed Alrahmawy,
authors use IWDs algorithm to distribute tasks to VMs and Alaa Eldin Reyad. “An extended Intelligent Water Drops algorithm for
execute them accordingly. In this way, the proposed system is workflow scheduling in cloud computing environment” Science Direct
Published by Elsevier; 2017
expected to perform better for task scheduling in Cloud for
[7] Dr. Sudha Sadhasivam, R. Jayarani, Dr. N. Nagaveni, R. Vasanth Ram
both homogeneous and heterogeneous Virtual Machines. “ Design and Implementation of an efficient T wo- level Scheduler for
Cloud Computing Environment” In Proceedings of International
Conference on Advances in Recent Technologies in Communication
and Computing, 2009
[8] Shah-Hosseini H. Problem solving by intelligent water drops. In: IEEE
congress on evolutionary computation; 2007
[9] Dadaneh BZ, Markid HY, Zakerolhosseini A. Graph coloring using
Intelligent water drops algorithm. In: 2015 23rd Iranian conference on
electrical engineering (ICEE), IEEE; 2015
[10] Kamkar I, Akbarzadeh-T M-R, Yaghoobi M. Intelligent water drops a
new optimization algorithm for solving the vehicle routing problem. In:
2010 IEEE international conference on systems man and cybernetics
(SMC), IEEE; 2010.
[11] Noferesti S, Shah-Hosseini H. A hybrid algorithm for solving steiner
tree problem. Int J Comput Appl 2012;41(5):14 –20.
[12] Shah-Hosseini H. Intelligent water drops algorithm: a new
optimization method for solving the multiple knapsack problem. Int J
Intell Comput Cybern 2008;1(2):193–21
[13] Rayapudi SR. An intelligent water drop algorithm for solving
economic load dispatch problem. Int J Electr Electron Eng
2011;5(2):43–9.
[14] Duan H, Liu S, Lei X. Air robot path planning based on intelligent
Fig. 2: Proposed System Model
water drops optimization. In: IEEE international joint conference on
neural networks, 2008. IJCNN 2008. (IEEE world congress on
computational intelligence), IEEE; 2008
[15] Kayvanfar V, Zandieh M, T eymourian E. An intelligent water drop
VI. CONCLUSION algorithm to identical parallel machine scheduling with controllable
Cloud resources are managed and shared by various cloud processing times: a just -in-time approach. Comput Appl Math 2015:1–
26.
users by using the task scheduling algorithm. This paper has [16] Chen W-N, Zhang J. An ant colony optimization approach to a grid
briefly discussed about various existing task scheduling workflow scheduling problem with various QoS requirements. IEEE
algorithms and compared them based on various cloud T rans Syst Man Cybern Part C: Appl Rev 2009;39(1):29 –43.
parameters, their advantages, disadvantages and future works. [17] Pandey S et al. A particle swarm optimization-based heuristic for
scheduling workflow applications in cloud computing environments.
From the survey, it is very clear that, there are many areas In: 2010 24th IEEE international conference on advanced information
where further attention is required e.g. there is a need to make networking and applications (AINA), IEEE; 2010.

978-1-5386-9439-8/19/$31.00 ©2019 IEEE 44

 Proceedings of the Third International Conference on Trends in Electronics and Informatics (ICOEI 2019)
IEEE Xplore Part Number: CFP19J32-ART; ISBN: 978-1-5386-9439-8

[18] Hu X-H et al. An IPSO algorithm for grid task scheduling based on
satisfaction rate. In: International conference on intelligent human -
machine systems and cybernetics, 2009. IHMSC’09, IEEE; 2009.
[19] Wu Z et al. A market -oriented hierarchical scheduling strategy in cloud
workflow systems. J Supercomput 2013;63(1):256–93.
[20] Lin C, Lu S. Scheduling scientific workflows elastically for cloud
computing. In: 2011 IEEE international conference on cloud
computing (CLOUD), IEEE; 2011.
[21] Xu M et al. A multiple QoS constrained scheduling strategy of multiple
workflows for cloud computing. In: 2009 IEEE International
Symposium on parallel and distributed processing with applications,
IEEE; 2009.
[22] Wu Z et al. A revised discrete particle swarm optimization for cloud
workflow scheduling. In: 2010 international co nference on
computational intelligence and security (CIS), IEEE; 2010
[23] Bilgaiyan S, Sagnika S, Das M. Workflow scheduling in cloud
computing environment using cat swarm optimization. In: 2014 IEEE
international advance computing conference (IACC), IEEE; 201 4.
[24] M. B. Yougita Chawla, "A study on Scheduling Methods in Cloud
Computing," International Journal of Emerging T rends & T echnology
in Computer Science (IJET T CS), vol. 1, no. 3, October 2012.
[25] Lan Wang and Erol Gelenbe, Adaptive Dispatching of T asks in the
Cloud, DOI 10.1109/TCC.2015.2474406, IEEE Transactions on Cloud
Computing.
[26] U. Bhoi, P. N. Ramanuj, and W. S. Email, “Enhanced Max -min T ask
Scheduling Algorithm in Cloud Computing,” vol. 2, no. 4, pp. 259 –
264, 2013.
[27] U. Bhoi, P. N. Ramanuj, and W. S. Email, “ Enhanced Max-min T ask
Scheduling Algorithm in Cloud Computing,” vol. 2, no. 4, pp. 259 –
264, 2013.
[28] Meng X, Liu Y, Gao X, Zhang H (2014) A new bio -inspired
algorithm: chicken swarm optimization. In: International Conference
in Swarm Intelligence, pp 86–94
[29] Rani E, Kaur H (2017) Efficient load balancing task scheduling in
cloud computing using raven roosting optimization algorithm. Int J
Adv Res Comput Sci 8(5):2419–2424
[30] J. T aheri, Y. Lee Ch, A.Y. Zomaya, H. Jay Siegel,“A Bee Colony
based optimization approach for simultaneous job scheduling and data
replication in grid environments”,Computers & Operations
Research,Vol. 40, 2013, pp. 1564–1578.
[31] Auila S, Amandeep U (2015) T ask scheduling in cloud using hybrid
cuckoo algorithm. Int J Comput Netw. Appl 2(3):144 -150
[32] Yang, X.S.: Bat algorithm: literature review and applications.
International Journal of Bio-inspired Computation 5(3), 141–149
(2013)
[33] Venkatesan, R., Ramalakshmi, K., Latha, D.P.P. Review on various
approaches for resource management and job scheduling in Grid
environment (2018) Journal of Computational and T heoretical
Nanoscience, 15, pp. 2682-2688.

978-1-5386-9439-8/19/$31.00 ©2019 IEEE 45