SCIENTIFIC WORKFLOW TASK BY

SCHEDULING IN CLOUD
ABSTRACT: Scientific workflows are critical in processing adaptability to real-time changes in cloud resource
complex data and computations across diverse domains, availability.
including bioinformatics, climate modelling, and physics
simulations. As data and computational requirements The main scope of this project is to provide
increase, cloud computing offers scalable resources for communication continuously by checking the energy at
executing these workflows efficiently. However, optimizing every communication and also to find out the intrusion
task scheduling within scientific workflows presents that is happening inside the network. The main
significant challenges, particularly in a cloud environment objectives of the projects are as follows:To find out
with dynamic and heterogeneous resources. Existing
early attacks, to minimize packet loss, More
scheduling approaches often struggle with issues such as
high cost, inefficiency, and resource contention. This paper
throughputs, to reduce time consumption, continuous
proposes an enhanced scheduling model tailored for energy check up of all nodes to avoid node failure.
scientific workflows in the cloud, aimed at minimizing
execution time and cost while maximizing resource
utilization. The proposed model incorporates dynamic load RELATED WORK
balancing, task prioritization, and resource adaptability,
resulting in improved efficiency and cost-effectiveness. Literature evaluation is an important step inside the
Experimental results demonstrate that the proposed software program development manner.
approach outperforms traditional scheduling models,
making it a valuable solution for scientific workflows in Jun Yan; Zhifeng Liu; Tao Zhang; Yueze ZhangWith
cloud-based environments. the development of intelligent manufacturing
technology, automatic guided vehicle (AGV) has been
KEYWORD: Cloud, Scheduling task, Data monitors,
widely used in workshop logistics transportation.
Hestrogeneous resources, Scientific workflow.
However, the complex and changeable production
INTRODUCTION process requires efficient and dynamic transportation
and distribution. In this paper, based on the results of
Scientific workflows are complex sets of process scheduling, a method of independent decision-
computational tasks organized to process and analyze making transportation process is studied. A
large-scale datasets in various scientific domains. As transportation strategy training method with breakpoint
scientific research increasingly relies on computational continuation and hierarchical feedback is proposed
power, cloud computing has emerged as an ideal based on deep Q-network (DQN). Transportation
platform due to its scalable and flexible resource scheduling can be quickly decided and adjusted to
offerings. Cloud computing allows researchers to adapt to dynamic and changeable orders. The method
execute workflows at scale, meeting the high demands is applied to the FJSP problem, and the data
of modern scientific computation. However, efficient experiments show the effectiveness of the method.[2]
task scheduling within scientific workflows remains a
challenge in cloud environments. Tasks in a scientific Victor Toporkov; Dmitry Yemelyanov; Artem
workflow are often interdependent, varying in Bulkhak. In this work we study a promising approach
execution time and resource requirements. Scheduling for efficient online scheduling of job-flows in high
these tasks optimally requires balancing resources performance and distributed parallel computing. The
while minimizing time and cost—an objective that is majority of job-flow optimization approaches,
complex due to the variability in cloud resources and including backfilling and microscheduling, require
the dynamic nature of scientific workloads. apriori knowledge of a full job queue to make the
optimization decisions. In a more general scenario
Traditional scheduling models struggle to meet these when user jobs are submitted individually, the
requirements, often leading to high costs, inefficient resources selection and allocation should be performed
resource use, and delays. For scientific applications immediately in the online mode. In this work we
that demand timely results, such inefficiencies can consider a neural network prototype model trained to
hinder research progress. Thus, a new approach to task perform online optimization decisions based on a
scheduling in scientific workflows is essential to known optimal solution. For this purpose, we designed
leverage cloud computing’s full potential. The MLAK algorithm which implements 0-1 knapsack
proposed scheduling model in this study addresses problem based on the apriori unknown utility function.
these challenges by optimizing resource utilization, In a dedicated simulation experiments with different
balancing task loads dynamically, and incorporating utility functions MLAK provides resources selection

Cisco Confidential
efficiency comparable to a classical greedy algorithm. In computational science, particle swarm
The interdisciplinary nature of the considered problem optimization (PSO) is a computational method
and its solution (including elements from the areas of that optimizes a problem by iteratively trying to
optimization, scheduling algorithms, machine learning improve a candidate solution with regard to a given
and simulation) makes it a good study example in measure of quality. It solves a problem by having a
technical disciplines.[3] population of candidate solutions, here
dubbed particles, and moving these particles around in
Manish Chhabra; Gagandeep Arora. Efficient task the search-space according to simple mathematical
organization is essential to maximizing cloud formulae over the particle's position and velocity. Each
computing's performance. We provide a new machine particle's movement is influenced by its local best
learning (ML) approach to evaluate and rank task known position , but is also guided toward the best
scheduling algorithms based on their features. By known positions in the search-space, which are
employing GDrive and the Spacy English model for updated as better positions are found by other particles.
feature extraction, we identify and quantify important This is expected to move the swarm toward the best
phrases related to algorithm characteristics. We solutions.
establish the relative importance of each attribute by
giving these terms precedence based on our PSO is originally attributed to Kennedy, Eberhart and
methodology, particularly based on their frequency. Shi and was first intended for simulating social
We then combine these priorities to get priority scores behavior, as a stylized representation of the movement
for each algorithm, which reveal the algorithms' of organisms in a bird flock or fish school. The
performance potential. An analysis of the priority algorithm was simplified, and it was observed to be
ratings can be used to determine the best work performing optimization. The book by Kennedy and
scheduling algorithm. The X- Y plot that displays Eberhart describes many philosophical aspects of PSO
these ratings makes it easier for people to comprehend and swarm intelligence. An extensive survey of PSO
and contrast the various algorithms. With the help of applications is made by Poli
our innovative method, users-whether personal or
business-can make well-informed decisions that PROPOSED SYSYTEM
maximize cloud resource use and boost overall Atomic Orbital Search (AOS) Algorithm, AOS is a
productivity. In order to optimize cloud computing computational method used in quantum chemistry for
efficiency, this research paper suggests a machine the calculation of the electronic structure of atoms and
learning-driven idea backed by feature to follow molecules. The AOS algorithm is based on the solution
navigate several work algorithms for scheduling. In of the Schrödinger equation, which describes the
cloud environments, artificial intelligence's power behavior of electrons in a system. The Schrödinger
greatly aids in achieving greater resource efficiency equation is a partial differential equation, which cannot
and enhanced performance.[4] be solved exactly for most systems of interest.
Gu Wenbin; Li Yuxin; Wang Yi. To deal with the Therefore, approximate methods are used to solve the
fastly increasing energy consumption and energy costs equation, and the AOS algorithm is one such method.
in the manufacturing process, manufacturing The AOS algorithm starts by assuming that the
companies are forced to search some methods to electronic wave function of the system can be
reduce energy cost without affecting the yield of their represented by a linear combination of atomic orbitals
products or sacrificing quality. In this paper, an (LCAO), which are the solutions of the Schrödinger
extended job shop scheduling problem, where equation for isolated atoms. These atomic orbitals are
machines can work at different speeds with different also referred to as basic functions. The coefficients of
energy consumption, is modified. A serial of analytical the LCAO are then determined by solving the secular
functions are developed to denote the relationship equation, which is a matrix equation. The secular
between Energy-efficiency and Makespan. And a equation is derived from the Schrödinger equation by
improved particle swarm algorithm, inspired by the inserting the LCAO ansatz into the equation and
hormone mechanism, is developed to solve the energy- applying the variational principle. The variational
efficient job shop scheduling problem. In term of principle states that the energy of the system is
practical application aspect, the result from minimized when the wave function is chosen to be a
computational experiments indicate that durations of member of a set of functions called the trial functions.
processing periods and processing speed of machines The AOS algorithm can be divided into two steps: the
have great effect on energy-efficiency scheduling initialization step and the self-consistency step. In the
solution, and our algorithm could solve instances with initialization step, the atomic orbitals are chosen, and
a good solution quality.[5] the secular equation is solved for the first time. In the
EXISTING SYSTEM self-consistency step, the secular equation is solved
iteratively until the coefficients of the LCAO converge
Particle Swarm Optimization (PSO): to a certain threshold. The choice of atomic orbitals is
important in the AOS algorithm, as it can affect the

Cisco Confidential
accuracy and efficiency of the calculation. The most Packet scanner: After catching the packet we use
common choice of atomic orbitals is Gaussian type packet scanner for the purpose of scanning the packet.
orbitals (GTO), which are functions that are defined by Packet scanning is the important part of our system.
a Gaussian function multiplied by a polynomial. GTOs
are chosen due to their mathematical simplicity and Packet analyzer: Packet analyzer is also known as
flexibility. They are also easy to integrate and can be Packet sniffer. As data streams flow across the
used to represent both valence and core orbitals. network sniffer capture each packet and if needed
Another common choice is Slater type orbitals (STO), decodes the packet, raw data showing the values of
which are defined by a radial function multiplied by a various fields in the packet and analyzes its content
polynomial. STOs are more physically motivated than accord- ing to specification.
GTOs, but they are less flexible. Labeling: Labeling is used for defining the
The AOS algorithm is a powerful tool for the corresponding packet.
calculation of the electronic structure of atoms and Training model: Training model contain set of trained
molecules. It has been used to study a wide range of data set which used to detect attack in the packet.
systems, including small molecules, large
biomolecules, and solids. The method is relatively easy Prediction: Based on the training model we make the
to implement and is well suited for parallel computing. prediction that the packet is normal packet or abnormal
However, the AOS algorithm has some limitations. packet.
One of the main limitations is that it is based on the
Output: Based on the prediction (i.e. normal or
LCAO approximation, which may not be accurate for
abnormal) we generate output in the form for abnormal
systems with a large electron correlation. Another
packet.
limitation is that the AOS algorithm is computationally
intensive, which can make the calculation time- 4.2 ALGORITHM OF PROPOSED WORK
consuming for large systems. In conclusion, the atomic
orbital search algorithm is a powerful computational i. Dataset is divided into small overlapping or either
method used for the calculation of the electronic non-overlapping blocks.
structure of atoms and molecules. The method is based ii. Extract the features using traditional techniques.
on the solution of the Schrödinger equation and the
LCAO approximation, and it involves the choice of iii. Extracted feature values corresponding to each key
atomic orbitals, the solution of the secular equation, point are stored in matrix.
and the self-consistency step.
iv. Apply sorting techniques to get similar features that
The comparison of the quantum staircase draws a lie in nearness.
parallel to the electron’s progression across distinct
orbitals, achieved through energy level alterations. In v. Introduce shift vector concept to find key point with
cases where an electron assimilates energy beneath its similar shifting.
binding threshold, it advances to a higher energy level vi. Use the counter vector to count the occurrence
in the outer orbital. Conversely, if an electron absorbs same shifting key point and set the counter to 1.
energy surpassing its binding limit, it shifts to a lower
energy tier within the inner orbital. vii. Similar regions are identified with the help of
threshold value above steps are used.
Advantages
 Scalability.
 Accuracy. FLOW CHART
 High Detection Speed.
SYSTEM IMPLEMENTATION
Incoming packet: In our project directly connected to
network (we capture online packet) and transfer to
packet scanner.
Packet Capture: In Real time networking many
packets are transferred from source to destination. In
this module live packets from the network are captured
and passed to the pre-processor module. According to
the concept of machine learning the data is grouped or
clustered based on its features or characteristics for e.g.
protocols used by the packets.

Cisco Confidential
  Modeling, observing, and simulating the work
activities of the current system.
The goal of the feasibility study is to consider
alternative information systems solutions, evaluate
their feasibility, and propose the alternative most
suitable to the organization. The feasibility of a
proposed solution is evaluated in terms of its
components. These components are:
 ECONOMICAL FEASIBILITY
 TECHNICAL FEASIBILITY
 SOCIAL FEASIBILITY
 OPERATIONAL FEASIBILITY
2. ECONOMICAL FEASIBILITY:
The This study is carried out to check the economic
impact that the system will have on the organization.
The amount of fund that the company can pour into the
research and development of the system is limited. The
expenditures must be justified. Thus the developed
system as well within the budget and this was achieved
because most of the technologies used are freely
available. Only the customized products had to be
purchased.
SYSTEM STUDY 3. TECHNICAL FEASIBILITY:
1. Feasibility Study: This study is carried out to check the technical
The The feasibility of the project is analyzed in this feasibility, that is, the technical requirements of the
phase and business proposal is put forth with a very system. Any system developed must not have a high
general plan for the project and some cost estimates. demand on the available technical resources. This will
During system analysis the feasibility study of the lead to high demands on the available technical
proposed system is to be carried out. This is to ensure resources. This will lead to high demands being placed
that the proposed system is not a burden to the on the client. The developed system must have a
company. For feasibility analysis, some understanding modest requirement, as only minimal or null changes
of the major requirements for the system is essential. are required for implementing this system.

The feasibility study investigates the problem and the

information needs of the stakeholders. It seeks to RESULT & FUTURE SCOPE
determine the resources required to provide an
information systems solution, the cost and benefits of Sample In conclusion, resource optimization
such a solution, and the feasibility of such a solution. techniques based on job scheduling play a crucial role
The analyst conducting the study gathers information in maximizing the efficiency and performance of cloud
using a variety of methods, the most popular of which computing systems. By effectively managing and
are: allocating resources, these techniques help minimize
costs, reduce energy consumption, and improve overall
 Developing and administering questionnaires system throughput. One of the key findings is that
to interested stakeholders, such as potential intelligent job scheduling algorithms, such as genetic
users of the information system. algorithms, ant colony optimization, and particle
 Observing or monitoring users of the current swarm optimization, can significantly enhance
system to determine their needs as well as their resource allocation in cloud environments. These
satisfaction and dissatisfaction with the current algorithms consider various factors, such as job
system. priority, resource availability, andtask dependencies, to
make informed decisions and optimize resource
 Collecting, examining, and analyzing utilization. Moreover, dynamic resource provisioning
documents, reports, layouts, procedures, techniques enable the system to adapt to changing
manuals, and any other documentation relating workload conditions. By continuously monitoring
to the operations of the current system. resource usage and workload patterns, these techniques

Cisco Confidential
can scale resources up or down in real-time, ensuring Apache 2 and IIS 10.0 Web Servers,” in 2018
efficient resource allocation and meeting service level International Conference on Advanced Science and
agreements (SLAs). Furthermore, the integration of Engineering (ICOASE), 2018, pp. 156–161.
machine learning and artificial intelligence techniques
in job scheduling can bring further improvements. By 7. R. R. Zebari, S. R. Zeebaree, K. Jacksi, and H. M.
leveraging historical data and learning patterns, these Shukur, “E-Business Requirements For Flexibility And
techniques can predict future resource demands and Implementation Enterprise System: A Review.”
optimize scheduling decisions accordingly. This helps International Journal of Scientific & Technology
in proactive resource management and avoiding Research (IJSTR), vol. 8, no. 11, pp. 655-660, 2019.
potential bottlenecks or performance degradation. 8. S. R. Zeebaree, K. Jacksi, and R. R. Zebari, “Impact
Overall, resource optimization techniques based on job analysis of SYN flood DDoS attack on HAProxy and
scheduling in cloud computing have shown great NLB cluster-based web servers,” Indones. J. Electr.
potential in achieving efficient resource utilization, Eng. Comput. Sci., vol. 19, no. 1, pp. 510–517, 2020.
cost reduction, and improved system performance
9. S. R. Zeebaree, R. R. Zebari, and K. Jacksi,
“Performance analysis of IIS10. 0 and Apache2
FUTURE SCOPE Cluster-based Web Servers under SYN DDoS Attack,”
TEST Engineering & Management, vol. 83, no. March
Future research should consider other machine - April 2020, 5854 - 5863, 2020.
learning algorithms to ascertain more efficient ways to
perform the classification technique on the datasets. It 10. S. R. Zeebaree, R. R. Zebari, K. Jacksi, and D. A.
is recommended that further research should be carry Hasan, “Security Approaches For Integrated Enterprise
out on other parameters that can improve the accuracy Systems Performance: A Review.” International
of detection Journal of Scientific & Technology Research (IJSTR)
vol. 8, no. 12, pp. 2485-2489, 2019.
11. R. Zebari, A. Abdulazeez, D. Zeebaree, D. Zebari,
and J. Saeed, “A Comprehensive Review of
Dimensionality Reduction Techniques for Feature
REFERENCES Selection and Feature Extraction,” J. Appl. Sci.
1. A. T. Saraswathi, Y. R. A. Kalaashri, and S. Technol. Trends, vol. 1, no. 2, pp. 56–70, May 2020,
Padmavathi, “Dynamic Resource Allocation Scheme doi: 10.38094/jastt1224.
in Cloud Computing,” Procedia Comput. Sci., vol. 47, 12. A. A. S. Farrag, S. A. Mahmoud, and E. S. M. El-
pp. 30–36, Jan. 2015, doi: Horbaty, “Intelligent cloud algorithms for load
10.1016/j.procs.2015.03.180. balancing problems: A survey,” presented at the 2015
2. P. Priyadarshinee, R. D. Raut, M. K. Jha, and B. B. IEEE Seventh International Conference on Intelligent
Gardas, “Understanding and predicting the Computing and Information Systems (ICICIS), Dec.
determinants of cloud computing adoption: A two 2015, pp. 210–216, doi:
staged hybrid SEM - Neural networks approach,” 10.1109/IntelCIS.2015.7397223.
Comput. Hum. Behav., vol. 76, pp. 341–362, Nov. 13. R. Rengasamy and M. Chidambaram, “A Novel
2017, doi: 10.1016/j.chb.2017.07.027. Predictive Resource Allocation Framework for Cloud
3. S. Vakilinia, M. M. Ali, and D. Qiu, “Modeling of Computing,” in 2019 5th International Conference on
the resource allocation in cloud computing centers,” Advanced Computing Communication Systems
Comput. Netw., vol. 91, no. Supplement C, pp. 453– (ICACCS), Mar. 2019, pp. 118–122, doi:
470, Nov. 2015, doi: 10.1016/j.comnet.2015.08.030. 10.1109/ICACCS.2019.8728526.

4. O. Alzakholi, L. Haji, H. Shukur, R. Zebari, S. 14. T. Deepika and A. N. Rao, “Active resource
Abas, and M. Sadeeq, “Comparison Among Cloud provision in cloud computing through virtualization,”
Technologies and Cloud Performance,” J. Appl. Sci. in 2014 IEEE International Conference on
Technol. Trends, vol. 1, no. 2, pp. 40–47, Apr. 2020, Computational Intelligence and Computing Research,
doi: 10.38094/jastt1219. Dec. 2014, pp. 1–4, doi:
10.1109/ICCIC.2014.7238373.
5. D. Kesavaraja and A. Shenbagavalli, “QoE
enhancement in cloud virtual machine allocation using 15.O. M. Ahmed and W. M. Abduallah, “A Review on
Eagle strategy of hybrid krill herd optimization,” J. Recent Steganography Techniques in Cloud
Parallel Distrib. Comput., Sep. 2017, doi: Computing,” Acad. J. Nawroz Univ., vol. 6, no. 3, pp.
10.1016/j.jpdc.2017.08.015. 106–111, 2017.

6. R. R. Zebari, S. R. Zeebaree, and K. Jacksi, “Impact

Analysis of HTTP and SYN Flood DDoS Attacks on

Cisco Confidential
16. O. M. Ahmed and A. B. Sallow, “Android security:
a review,” Acad. J. Nawroz Univ., vol. 6, no. 3, pp.
135–140, 2017.
17. K. Zaki and H. Saad, “Adoption of Cloud Human
Resource Information System in Egyptian Hotels: An
Experimental Design Research,” Int. J. Herit. Tour.
Hosp., vol. 12, no. 1, pp. 233–245, Mar. 2018.
18. N. Harki, A. Ahmed, and L. Haji, “CPU
Scheduling Techniques: A Review on Novel
Approaches Strategy and Performance Assessment,” J.
Appl. Sci. Technol. Trends, vol. 1, no. 2, pp. 48–55,
2020.

Cisco Confidential