A review of cloud
based applications and efficient resource utilization and allocation
Amita Boral1*, Dr. Kishan
Kumar2
1 Research Scholar, Shri Krishna University,
Chhatarpur, M.P.
ouriginal.sku@gmail.com
2 Professor, Shri Krishna
University, Chhatarpur, M.P.
Abstract- Cloud-based
applications have revolutionized computing by offering scalable, flexible, and
cost-effective solutions across diverse industries. These applications rely on
efficient resource utilization and allocation to ensure optimal performance,
minimize costs, and enhance user satisfaction. This review explores the
evolution, architecture, and benefits of cloud-based applications, emphasizing
the challenges associated with resource management. Key strategies for
improving resource allocation, including virtualization, load balancing, task
scheduling, and dynamic resource provisioning, are analyzed. Additionally,
emerging technologies such as artificial intelligence, machine learning, and
edge computing are examined for their role in optimizing resource allocation.
Through a comprehensive assessment of existing studies and methodologies, this
review highlights the importance of innovative solutions for addressing
inefficiencies, managing workloads, and reducing latency.
Keywords- Cloud Computing,
Benefits Machine Learning, Resource Allocation, Virtualization
INTRODUCTION
The modern era of computing is characterized by the
widespread use of cloud-based applications and services, providing businesses
and individuals with unprecedented scalability, flexibility, and accessibility.
The ever-increasing need for these cloud-based applications, however, has
presented additional difficulties, such as the optimal allocation of cloud
resources & enhancement of application performance. By optimizing
cloud application performance & assuring optimal resource use, Machine
Learning (ML) approaches have emerged as a potent tool for tackling these issues.
Here, we will lay the groundwork for a more in-depth investigation of Machine
Learning's contribution to the betterment of cloud-based applications &
resource management by outlining the study's core goals, significance, and
significant issues.
CLOUD COMPUTING
The term "cloud computing" refers to a model
of computing that allows users to share resources (such as servers) and
applications over the internet, as well as the underlying platform. The phrase
"cloud" is a metaphor that has taken on an international connotation,
indicating anything that extends across the whole planet. Cloud computing
refers to both a platform & type of application. Servers, in the form
of either existing physical computers or virtualized versions of such machines,
can be provided, configured, & reconfigured by cloud platform services.
Conversely, online applications and web services are hosted by huge data
centers and dominating servers in what is known as "Cloud Computing,"
which explains how applications are expanded to be accessible through the
internet.
To maximize cloud performance and resource use, it
employs virtualization & load balancing techniques. In addition to this, it
employs technology such as web services, distributed computing, networking,
etc. on this context, the word "cloud" refers to a globally
distributed, configuration-free server on the cloud. People and major
corporations alike can take advantage of software and hardware services
provided by remote third parties through the cloud paradigm. This includes
things like online file storage, social media, email, & business
applications.
As a means of negotiating between cloud service providers and their clients,
SLAs (Service Level Agreements) include QoS factors like time limitations
and ensure that performance statistics are adhered to. By utilizing cloud
computing, SMEs can free themselves from the burden of managing, securing,
configuring, and selling their IT infrastructure.
APPLICATIONS OF CLOUD COMPUTING
There is a daily growth in the number of companies
engaged in cloud-based service trading. The following are some of the many
fields in which cloud computing can provide its services:
·
Testing and Development:
The cloud paradigm generates the availability of tools for application
building. In order to make application development, testing, and deployment
accessible to non-developers, it provides tools that are easy to configure,
thus removing the complexity. Online development frameworks reduce the need to
spend time and money on resource acquisition.
·
Backup and Disaster
Recovery: In the event of a breakdown, data and
services stored in the cloud are replicated for recovery purposes.
·
Large data storage:
Web-enabled interfaces make it easier to store large data sets in cloud
datacenters. Skilled teams handle storage management, security, privacy, and
consolidation tasks via the cloud, relieving people and companies of the burden
of worry about large investments and maintenance needs.
·
Software Library:
The cloud allows organizations to easily get software as needed, whenever they
need it, without getting caught up in the legal complexities of pirated
versions.
·
E-Commerce:
The advent of cloud computing in the worldwide market has recently boosted the
trading of commodities over the internet. Season, festival, and holiday
fluctuations affect the e-commerce industry's cloud traffic. The cloud is able
to automatically scale up during peak demand and down during off-peak periods
because of its proficiency in resource provisioning.
·
Website Hosting:
Developers' focus shifts from development to management when they engage in
hosting activities. By delegating maintenance duties to teams who specialize in
providing services, developers may rest easy while their apps are hosted on the
cloud.
·
Databases in the Cloud:
Developers now have access to complicated and expert-level solutions for tuning
their databases on the cloud. By moving their services to the cloud, service
providers are increasing the rate at which they generate income. Rackspace is
one company that has used this approach.
·
Electronic mail:
E-mail's move to the cloud has increased storage scalability, made client data
more secure, and protected privacy. Also, users no longer have to worry about
server outages.
SERVICES OF CLOUD COMPUTING
Cloud computing allows for the simultaneous provision
of numerous services to a huge user base. The SPI model, which is a component
of cloud computing, encapsulates the various services that the cloud provides.
As shown in figure 1, the cloud is viewed from the standpoint of the services
it offers and the deployment model it supports.
Figure 1: Cloud Computing perspective
Figure 2: Cloud Computing service models
SaaS, PaaS, and IaaS are the three main subsets
of cloud services, as seen in Figure 2. Following is a brief overview of the
three models:-
Software as a Service (SaaS):
Cloud computing eliminates the need for individuals & businesses to
purchase, maintain, and upgrade software by providing on-demand software
services through subscription. The provider can control who has access to the
SaaS app & how often it needs maintenance because it is hosted and operated
on their servers. The program is typically rented out on a per-use basis or
purchased as a subscription, with each user receiving their own license. Users
in this paradigm are not concerned with the underlying infrastructure or
platform; rather, they merely need to access the service as a web application.
The SaaS family of services is enriched by applications like online games,
social media sites, and office software.
Platform as a Service (PaaS):
Platform as a service, or PaaS, is a comprehensive set of tools that developers
may use to build and launch their apps in the cloud, whether it's a public or
private one. By doing away with the complexity of managing individual software
& hardware components, it allows an organization to take use of critical
middleware services. PaaS examples include Engine Yard, Google App
Engine, Windows Azure, and Force.com.
Infrastructure as a Service (IaaS):
By using the IaaS model, a company can contract out the management of its
storage, hardware resources, servers, & networking components, among other
things, to third parties. It is the service provider's responsibility to house
and maintain the equipment, as he owns it. The customer makes a one-time
payment. A safe, standardized, and extensible foundation is the bedrock upon
which infrastructure services are constructed. High availability &
elasticity of resources necessitate virtualization and some degree of
infrastructure redundancy. The foundation of operating services in virtualized
environments is server virtualization, most commonly via VMware or XEN.
Software automation should be used to easily provide and de-provision these
services.
CLOUD COMPUTING ADVANTAGES
Despite the fact that cloud computing resources are
distributed across global boundaries, the paradigm nonetheless provides its
entities with a wide range of benefits. The following are some of the
advantages provided:-
·
Cost Reduction:
Prominent corporations & people have adopted the cloud paradigm due to the
cloud's ability to consolidate infrastructure. The user no longer has to worry
about making a financial commitment to acquire hardware resources, software
licenses, and the necessary infrastructure to integrate them. In addition to
avoiding the first setback that occurred when work was about to begin as a
result of infrastructure setup procedures, this method has gone a step further.
·
Software Upgrades:
As a result of more varied computing demands, service providers with more
advanced capabilities are moving their operations to the cloud, relieving users
of the responsibility of keeping their software up-to-date.
·
Application selection
flexibility: The cloud enables users to build and
utilize bespoke services from a variety of vendors, each with their own unique
set of skills that can handle complex request compositions.
·
Resilient Computing:
Services & data are mirrored over the cloud and made accessible globally to
lessen the effect of calamities.
·
Service Focused on Usability:
Metered capabilities take advantage of paying for a service only when it is
used. Consequently, the strategy aims to free the user from the responsibility
of upkeep. In addition, it prevents spending that may have gone toward
equipment purchases.
TYPES OF MACHINE LEARNING
The field of artificial
intelligence (AI) recognized as ML is concerned with teaching computers to
learn from and make judgments based on data on their own, without being
specifically taught to do so. To get better at a task, machine learning
algorithms look for patterns in data and draw conclusions about the task based
on those patterns.
Different machine learning
technique include
·
Supervised Machine
Learning- Supervised ML algorithms are utilized to
make predictions about the future by learning from data sets that have been
labeled & mapped to certain target values. The primary responsibility of a
supervised ML algorithm is to analyze the input data & assign a proper
classification to it. Only through training on a large, well labeled dataset
with well-defined classes would such a distribution be possible. The two types
of ML problems that a supervised ML algorithm is able to address are
classification & regression. Classification is a solution method used when
the underlying problem has a binary (yes/no) target variable. On the other
hand, Regression ML methods are used to handle issues where the target variable
is not categorized but continuous.
·
Unsupervised Machine
Learning- The ML model is trained by unsupervised
ML techniques using datasets that are neither labeled nor categorized.
Unsupervised ML algorithms examine a huge dataset to discover and learn data
insights such as patterns, classifications, and categories without human
supervision. Two types of unsupervised ML are Clustering and Association.
Algorithms based on clustering divide data into groups based on their shared
features. Association-based algorithms, on the other hand, seek out connections
between pieces of information that naturally belong together.
·
Semi-Supervised Machine
Learning- It is designed to compensate for the
shortcomings of both supervised and unsupervised ML methods. In semi supervised
learning, the ML model is trained using both labeled and unlabeled datasets.
·
Reinforcement Machine
Learning- Reinforcement ML makes use of a learning
paradigm centered on the use of feedback. The agent is not trained on any
supervised datasets and instead is given positive or negative reinforcement for
making the right or wrong choices. In Figure 3 we see how various machine
learning methods can be categorized.
Figure 3: Algorithm ML classification
QUALITY OF SERVICE (QOS) IN A CLOUD COMPUTING
QoS is used to define the procedures and tools used in
cloud computing to guarantee that data & programs will function as expected
under all conditions. Quality of service, or QoS, is a crucial component of
cloud computing because it provides a mechanism for controlling and
consistently delivering the standard of service expected by cloud customers.
Key Qualitative Service in the Cloud Features:
·
Service Level
Agreements (SLAs): SLAs are the
starting point for many QoS arrangements between cloud providers and their
clients. SLAs define the expected levels of service, including factors like
uptime, response times, and data availability. They set the baseline for QoS.
·
Resource
Allocation: QoS involves allocating resources
within the cloud infrastructure to ensure that each service or application gets
the necessary computational, storage, and network resources. This includes
dynamic resource allocation to handle varying workloads.
·
Performance
Monitoring: Cloud providers use monitoring tools
and techniques to track the performance of services and applications. These
tools can detect issues such as network congestion, server load, or latency.
·
Load
Balancing: Load balancing is a QoS mechanism that
evenly distributes network traffic and requests across multiple servers. This
ensures that no single server is overwhelmed and that all users experience
consistent performance.
·
Security
& Privacy: QoS in the cloud extends to security
& privacy measures. Protecting data and ensuring secure access to services
is essential for maintaining a high level of service quality.
·
Fault Tolerance:
QoS mechanisms should include fault tolerance and redundancy. Redundant systems
and data backups help ensure that services remain available even if individual
components fail.
·
Scalability:
Cloud services should be able to scale up or down as demand changes. QoS
measures must ensure that this can happen seamlessly, without impacting service
quality.
·
Network
Quality: Network QoS involves managing bandwidth
and minimizing latency. Techniques like Quality of Service (QoS) settings and
Content Delivery Networks (CDNs) are used to optimize network performance.
·
Prioritization:
Some applications or users may require higher QoS levels than others.
Prioritization mechanisms help ensure that critical services receive the
resources they need.
·
Dynamic
Adaptation: Cloud systems should adapt to
changing conditions. For instance, if a service experiences high demand, it
should be able to allocate more resources on the fly to maintain performance.
QoS in cloud computing is essential for businesses and
organizations that rely on the cloud for their operations. It ensures that
cloud services meet their specific needs and expectations, and it provides a
framework for monitoring and continuously improving service quality. Effective
QoS measures are a critical part of cloud service management and are essential
for building trust and reliability in cloud computing environments.
VIRTUAL MACHINE
In the same way that a physical machine (PM) runs
programs, VM does the same. There are two main types of VMs, distinguished
by how they are used and how closely they resemble physical machines. Figure.4
shows a virtual machine structure, which is detailed below.
Figure 4: Structure of VM
VM operates as a standalone process because its
sole purpose is to run a single application. Such VMs offer system
flexibility & ease of use while being well-suited to one or more
programming languages.
Virtual Machine Allocation
In a datacenter, VMs are in high demand from
cloud users for specific resources like storage, processing power, and memory.
Upon receiving this request from the client, the DC will assign the virtual
machine to a server and ensure that all critical resources are available on
that server. Depending on the server's capacity & VMs' resource needs,
a server can support a large number of VMs.
Since energy consumption has become a major expense
and environmental concern for server farms, it is critical to allocate
VMs to servers in a way that minimizes energy consumption. The CPU use of
the preset assets determines the position of a virtual machine when it is
chosen to relocate.
Virtual Machine Migration
Virtual Machine Migration, or VMM, refers to the
process of transferring virtual machines from one system to another. Tolerating
PM errors, balancing workloads, and reducing DC power usage are all made easier
with the VMM. It takes time for every task in the cloud to migrate among
virtual machines. The service delivery is not affected during the dynamic
migration of VMs in cloud data centers in order to fulfill customer
demands. New cloud events, such as workload balancing, online maintenance,
server consolidation, and so on, can be supported by the migration mechanism.
With VMM, cloud resource allocation is lightning quick, and operational costs
are dropping. In addition, the structure of excessive energy consumption is not
just the quantity of recording assets & power inefficiency of
equipment; it is the direct outcome of inefficient usage of these assets.
Virtual Machine Consolidation
Consolidating virtual machines (VMs) is an important
part of creating a dynamic cloud resource management system that uses less
energy. By assisting with the migration of virtual machines into physical
servers, the VMC allows for more efficient use of cloud resources with less
power consumption. However, bad QoS could result from processing multiple
VMs onto a single server. To tackle this, VMC algorithms are designed to
progressively evaluate the impact on QoS when choosing which virtual machines
to transfer. As shown in Figure.5, VMC moves the virtual machines with fewer
physical machines (VMs) than previously. At the same time, PMs without VMs can
be switched from an active (or "on") state to a less energy-intensive
(or "rest") state, like a sleep state, to minimise power consumption.
Figure 5: Architecture of VMC
DATACENTERS
RESOURCE
ALLOCATION
The four main components of resource management are
reporting, booking, allocating, and checking. In order to meet the demands of
clients, resource revelation identifies the best physical resources to create
virtual machines on. Out of all the coordinated physical resources, resource
planning selects the best one. In order to set up resources from a cloud
foundation, it actually identifies the physical resource that will house the
VM [Sriram Kailasam 2013]. The process of assigning resources eliminates
the need to select a specific resource for each task or assignment. Actually,
it means adjusting the work schedule to fit the selected cloud resource. Once
the accommodation of the occupation is complete, the resource is examined.
Figure 6: Resource allocations
Cloud-Based Energy Efficient Resource Allocation
In order to meet user requirements while improving
cost efficiency & reducing energy usage, etc., resource allocation is
primarily responsible for identifying & allocating resources to each
incoming user request. Figure.7 shows that schedulers have the option to either
ensure the static & underlying asset assignment at request arrival or to
distribute both static & dynamic assets in order to constantly oversee
assets, optimize, and check previous requests.
Figure 7: Resource Allocation in Cloud
Computing
Resource
Allocation Technique
A number of scheduling rules, such as the Global
scheduling policy, make use of the numerous aspects of the device in order to
allocate the work to the multiprocessor. Additionally, these policies manage
the performance of the system altogether. The following is a description &
listing of a crucial method of resource allocation:
•
Static Scheduling
Algorithm
•
Dynamic Scheduling
Algorithm
•
Heuristic Scheduling Algorithms
•
Opportunistic Load
Balancing
•
Min-Min technique
•
Max-Min technique
CONCLUSION
Efficient resource utilization and allocation are
critical for the sustained success of cloud-based applications. This review
demonstrates that effective resource management not only improves system
performance but also reduces operational costs and ensures reliability.
Techniques such as virtualization, task scheduling, and load balancing are
foundational, while the integration of AI and edge computing offers promising
advancements. However, challenges such as energy efficiency, workload
unpredictability, and security remain key concerns. Future research should
focus on developing adaptive and intelligent resource management frameworks
that leverage real-time analytics and predictive modeling. By addressing these
challenges, the cloud computing industry can achieve greater efficiency and
scalability, meeting the growing demands of users while fostering innovation
and sustainability.
References
1.
Al-Tous H. and Barhumi I.
(2016) ‘Resource allocation for multiple-sources singlerelay cooperative
communication OFDMA systems’, IEEE Transactions on Mobile Computing, Vo15, No.
4, pp. 964-981.
2.
B. Lawson, E. Smirni,
Power-aware resource allocation in high-end systems via online simulation, in
Proceedings of the19th Annual International Conference on Supercomputing,
Cambridge, USA 2005.
3.
Chen S. Wu J. and Lu Z.
(2012) ‘A cloud computing resource scheduling policy based on genetic algorithm
with multiple fitness’, In Computer and Information Technology (CIT), 2012 IEEE
12th International Conference on, pp. 177-184, IEEE, 2012
5.
Hitesh A. Ravani, Hitesh
A. Bheda, Vrunda J. Patel, “Genetic Algorithm Based Resource Scheduling
Technique in Cloud Computing”, International Journal of Advanced Research in
Computer Science and Management Studies, Volume 1, Issue 7, pp. 168-174, 2013.
6.
J. E. Haddad, M.
Manouvrier, G. Ramirez, and M. Rukoz, QoS-driven selection of web services for
transactional composition, InProc. 6th Int’l Conf. Web Services (ICWS’08),
2008, 653–660
7.
L. Shao, J. Zhang, Y. Wei,
J. Zhao, B. Xie, and H. Mei. Personalized QoS prediction for web services via
collaborative filtering, InProc. 5th Int’l Conf. Web Services (ICWS’07), 2007,
439–446.
8.
Kapil Bakshi, Cisco Cloud
Computing - Data Center Strategy, Architecture, and Solutions, Point of View
White Paper for U.S. Public Sector 1st Edition Cisco Systems, Inc., 2009, 1-16
9.
Ma Y.B. Jang S.H. and Lee
J.S. (2011) ‘Ontology-based resource management for cloud computing’, In Asian
Conference on Intelligent Information and Database Systems, pp. 343-352,
Springer, Berlin, Heidelberg, 2011.
10.
Sagar M.S. Singh B. and
Ahmad W. (2013) ‘Study on cloud computing resource allocation strategies’,
International Journal of Advance Research and Innovation, Vol. 1, No. 3, pp.
107-114.
11.
Reddy C. and Suchithra R.
(2016) ‘Virtual Machine Migration in Cloud Data Centers for Resource
management’, International Journal of Engineering and Computer Science, Vol. 5,
no 09, pp.18029-18034
12.
Vignesh V. Sendhil Kumar
K.S. and Jaisankar N. (2013) ‘Resource management and scheduling in cloud
environment’, International journal of scientific and research publications
Vol. 3, No. 6, pp. 1
13.
Wang S.C. Yan K.Q. Liao
W.P. and Wang S.S (2010) ‘Towards a load balancing in a three-level cloud
computing network’, In Computer Science and information technology (ICCSIT),
2010 3rd IEEE International Conference on, Vol. 1, pp. 108- 113, IEEE, 2010.
14.
Xie, R.; Jia, X.; Yang,
K.; Zhang, B. Energy saving virtual machine allocation in cloud computing.
Distributed Computing Systems Workshops (ICDCSW), 2013 IEEE 33rd International
Conference on: IEEE; 2013. p. 132-137.
15.
Yuan D. Yang Y. Liu X.
and Chen J. (2010) ‘A data placement strategy in scientific cloud workflows’,
Future Generation Computer Systems Vol. 26, No. 8, pp. 1200-1214.
16.
Z. Zheng and M. R. Lyu, A
distributed replication strategy evaluation and selection framework for fault
tolerant web services, InProc. 6th Int’l Conf. Web Services (ICWS’08), 2008,
145–152.