A LOAD BALANCING ALGORITHM FOR THE DATA

CENTRES TO OPTIMIZE CLOUD COMPUTING

APPLICATIONS

A PROJECT REPORT SUBMITTED TO

SRM INSTITUTE OF SCIENCE & TECHNOLOGY

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

AWARD OF THE DEGREE OF

MASTER OF COMPUTER APPLICATIONS

BY
MANONMANI M
REG NO. RA2232241010036

UNDER THE GUIDANCE OF

Mr. J. VENKATA SUBRAMANIAN, M.C.A., M.Phil., M.S., ̅̅̅̅̅̅̅̅̅
𝐏𝐡. 𝐃. ,

DEPARTMENT OF COMPUTER APPLICATIONS

COLLEGE OF SCIENCE AND HUMANITIES

SRM INSTITUTE OF SCIENCE & TECHNOLOGY

Kattankulathur – 603 203

Chennai, Tamil Nadu

April – 2024
 BONAFIDE CERTIFICATE

This is to certify that the project report titled “A LOAD BALANCING

ALGORITHM FOR THE DATA CENTERS TO OPTIMIZE CLOUD

COMPUTING APPLICATIONS” is a bonafide work carried out by MANONMANI

M (RA2232241010036) under my supervision for the award of the Degree of Master of

Computer Applications. To my knowledge the work reported herein is the original work

done by this student.

Mr. J. Venkata Subramanian Dr. S. Albert Antony Raj

Assistant Professor, Professor & Head,

Department of Computer Applications Department of Computer Applications

(GUIDE)

INTERNAL EXAMINER EXTERNAL EXAMINER

 DECLARATION OF ASSOCIATION OF RESEARCH PROJECT
WITH SUSTAINABLE DEVELOPMENT GOALS

This is to certify that the research project entitled “A LOAD BALANCING

ALGORITHM FOR THE DATA CENTRES TO OPTIMIZE CLOUD

COMPUTING APPLICATIONS” carried out by Mr. Deepan Vishwa S R under the

supervision of Mr. J. Venkata Subramanian in partial fulfilment of the requirement for

the award of Post Graduation program has been significantly or potentially associated

with SDG Goal No 09 (NINE) titled INDUSTRY AND INNOVATION AND

INFRASTRUCTURE

This study has clearly shown the extent to which its goals and objectives have

been met in terms of filling the research gaps, identifying needs, resolving problems, and

developing innovative solutions locally for achieving the above-mentioned SDG on a

National and/or on an international level.

SIGNATURE OF THE STUDENT GUIDE AND SUPERVISOR

HEAD OF THE DEPARTMENT

 ACKNOWLEDGEMENT

With profound gratitude to the ALMIGHTY, I take this chance to thank people
who helped me to complete this project.

I take this as a right opportunity to say THANKS to my parents who are there to
stand with me always with the words “YOU CAN”.

I am thankful to Dr.T.R. Paarivendhar, Chancellor, SRM Institute of Science &

Technology who gave us the platform to establish me to reach greater heights.

I earnestly thank Dr. A.Duraisamy, Dean, College of Science and Humanities,

SRM Institute of Science & Technology who always encourage me to do novel things.

I express my sincere thanks to Dr. S. Albert Antony Raj, M.Sc., M.Phil., Ph.D.,
Professor and Head, Department of Computer Applications, College of Science and
Humanities for his valuable guidance and support to execute all incline in learning.

It is my delight to thank my project guide Mr. J. Venkata Subramanian,

M.C.A., M.Phil., M.S., ̅̅̅̅̅̅̅
𝐏𝐡. 𝐃., Assistant Professor, Department of Computer
Applications for his help, support, encouragement, suggestions, and guidance throughout
the development phases of the project.

I convey my gratitude to all the faculty members of the department who extended
their support through valuable comments and suggestions during the reviews.

A great note of gratitude to friends and people who are known and unknown to
me who helped in carrying out this project work a successful one.

MANONMANI M
PROJECT CERTIFICATE
PLAGIARISM CERTIFICATE
 ABLE OF CONTENTS

CHAPTER .NO TITLE PAGE NO.

ABSTRACT
LIST OF FIGURES
LIST OF ABBREVIATIONS
1. INTRODUCTION
1.1 CLOUD DATA CENTER
1.2 VIRTUAL MACHINE
1.3 ENERGY CONSUMPTION
1.4 RESOURCE MANAGEMENT
1.5 OBJECTIVES
2 LITERATURE SURVEY
3 SYSTEM ANALYSIS
3.1 EXISTING SYSTEM
3.1.1 DRAWBACKS
3.2 PROPOSED SYSTEM
3.2.1 ADVANTAGES
3.3 FEASIBILITY STUDY
3.3.1 TECHNICAL FEASIBILITY
3.3.2 OPERATIONAL FEASIBILITY
3.3.3 ECONOMICAL FEASIBILITY
4 SYSTEM SPECIFICATION
4.1 HARDWARE CONFIGURATION
4.2 SOFTWARE SPECIFICATION
5 SOFTWARE DESCRIPTION
5.1 FRONT END
6 PROJECT DESCRIPTION
6.1 PROBLEM DEFINITION
6.2 MODULE DESCRIPTION
 6.3 SYSTEM FLOW DIAGRAM
6.4 INPUT DESIGN
6.5 OUTPUT DESIGN
7 SYSTEM TESTING AND IMPLEMENTATION
7.1 SYSTEM TESTING
7.2 SYSTEM IMPLEMENTATION
8 SYSTEM MAINTENANCE
8.1 CORRECTIVE MAINTENANCE
8.2 ADAPTIVE MAINTENANCE
8.3 PERFECTIVE MAINTENANCE
9 CONCLUSION AND FUTURE
ENHANCEMENT
10 APPENDICES
10.1 SOURCE CODE
10.2 SCREEN SHOTS
11 REFERENCES
 A LOAD BALANCING ALGORITHM FOR DATACENTERS TO OPTIMIZE CLOUD
COMPUTING APPLICATIONS

ABSTRACT

Virtual Machines (VMs) are assigned to hosts based on their immediate resource operation (e.g.,
hosts with the greatest available RAM) rather of their overall and long-term application.
Furthermore, the scheduling and placement procedures are often computationally intensive and
have an impact on the performance of stationed virtual machines. This work presents a Cloud
VM scheduling method that uses PSO fashion to optimize performance and record VMs by
analyzing previously executed VM resource operations over time through application scenarios.
The goal is to eliminate comparable performance declines because cloud operation operations,
such as virtual machine placement, have an impact on systems that were previously stationed.
Additionally, overloaded virtual machines (VMs) often take resources from nearby VMs, thus
the work optimizes the real CPU application of VMs. The outcomes suggest that our technique
improves traditional instant-based physical machine selection by learning system behavior and
adapting over time. The notion of VM scheduling using resource monitoring data taken from
previous resource utilizations (VMs). The PSO classifier reduces the number of physical
machines by four.
 CHAPTER 1

1. INTRODUCTION

In the dynamic landscape of cloud computing, the efficient management of resources within
virtualized data centers has become increasingly crucial. The escalating demand for computing
resources, coupled with the growing awareness of environmental sustainability, underscores the
need for innovative frameworks that not only optimize performance but also prioritize energy
efficiency. This is particularly relevant in the context of two-tier virtualized cloud data centers,
where the challenges of balancing resource allocation and minimizing energy consumption are
pronounced. In response to these challenges, an energy-aware host resource management
framework emerges as a promising solution. This framework aims to strike a delicate balance
between enhancing the overall performance of virtualized environments and mitigating the
environmental impact by intelligently allocating and managing computing resources. This
introduction sets the stage for a deeper exploration of the intricacies and benefits of such a
framework in the context of two-tier virtualized cloud data centers.

1.1 CLOUD DATA CENTER

In the era of rapid digital transformation, cloud data centers have emerged as the backbone of
modern computing infrastructure, revolutionizing the way businesses and individuals access and
manage data. These centers represent a pivotal shift from traditional, on-premises data storage and
processing to scalable and flexible computing environments hosted remotely. Cloud data centers
provide a vast array of services, ranging from storage and computation to networking and
analytics, enabling organizations to dynamically scale their resources based on demand. The
inherent advantages of scalability, cost efficiency, and accessibility have made cloud data centers
indispensable in today's interconnected world. As the demand for cloud services continues to soar,
understanding the intricacies of these data centers becomes essential for businesses and technology
enthusiasts alike. This introduction lays the groundwork for delving into the multifaceted world of
cloud data centers, exploring their architecture, functionalities, and the transformative impact they
wield in the realm of information technology.

1.2 VIRTUAL MACHINE

A virtual machine (VM) is a software-based emulation of a physical computer, enabling multiple
operating systems (OS) to run on a single physical machine. This technology allows for the
creation of isolated environments, known as virtualized instances or VMs, within which
applications and operating systems can operate independently of the underlying hardware. Key
components of virtual machines include a hypervisor or a Virtual Machine Monitor (VMM), which
is responsible for managing and allocating the physical resources of the host machine to the virtual
machines. There are two types of hypervisors: Type 1 (bare-metal) hypervisors run directly on the
hardware, while Type 2 (hosted) hypervisors run on top of an existing operating system. VMs are
widely used in various computing scenarios, such as server consolidation, testing and development
environments, and cloud computing. They provide benefits like improved resource utilization,
isolation, and the ability to run multiple operating systems on a single physical machine. Each
virtual machine has its own virtual CPU, memory, storage, and network interfaces, creating a
virtualized environment that operates independently of other VMs on the same host. In the context
of cloud computing, virtual machines are fundamental building blocks that enable users to deploy
and run applications in a flexible and scalable manner. Cloud providers offer VMs as on-demand
resources, allowing users to configure and deploy them based on their specific requirements
without the need to invest in or manage physical hardware. This flexibility and abstraction
contribute to the efficiency and agility of modern computing infrastructures.

1.3 ENERGY CONSUMPTION

In the contemporary landscape of technology and industry, energy consumption stands at the
forefront of global considerations. As societies increasingly rely on advanced technologies for
their daily operations, the demand for energy continues to escalate. From powering homes and
businesses to supporting the vast network of data centers that underpin our digital infrastructure,
the challenge lies not only in meeting these energy needs but also in doing so sustainably. The
environmental impact of energy consumption, particularly in the context of computing and data
processing, has become a significant concern. As the world transitions to more digital and
interconnected systems, understanding and addressing the energy implications of these
advancements are essential. This introduction sets the stage for an exploration of the complexities
surrounding energy consumption, emphasizing the critical need for innovative solutions and
frameworks that promote efficiency and sustainability across various sectors.
1.4 RESOURCE MANAGEMENT

Resource management stands as a linchpin in the orchestration of efficient and sustainable

operations across various domains, from business enterprises to computing environments. At its
core, resource management involves the judicious allocation, utilization, and optimization of
assets, be they human, financial, or technological. In the context of technology and computing,
effective resource management becomes especially critical. As organizations grapple with the
dynamic demands of the digital age, the ability to allocate computing resources judiciously,
ensuring optimal performance and responsiveness, becomes a strategic imperative. This
encompasses not only the allocation of physical assets such as servers and storage but also the
intelligent management of virtual resources in cloud computing environments. This introduction
sets the stage for a comprehensive exploration of resource management, shedding light on its
multifaceted nature and underscoring its pivotal role in achieving efficiency, resilience, and
sustainability in today's complex and interconnected systems.

1.5 OBJECTIVES
• Reduce energy consumption. The framework aims to reduce the overall energy consumption
of the cloud data center by consolidating containers on fewer hosts and turning off idle
hosts.

• The framework ensures that latency requirements for all containers are met by placing
containers on hosts that are close to their users.

• Improve performance: The framework aims to improve the overall performance of the cloud
data center by balancing the resource utilization of hosts.
 CHAPTER 2

2. LITERATURE REVIEW

2.1COMBINING VIRTUALIZATION AND CONTAINERIZATION TO SUPPORT

INTERACTIVE GAMES AND SIMULATIONS ON THE CLOUD

Sean C In order to address the challenges posed by the integration of game-based and virtual
software simulators into traditional networks, various organizations spanning the entertainment
industry, energy and financial sectors, military, and video gaming have turned to the powerful
capabilities of High-Performance Computing (HPC). The inherent capacity of HPC to handle
compute-intensive tasks makes it an attractive platform for running interactive simulations.
However, the focus of this work goes beyond the conventional use of HPC, aiming to explore the
feasibility of transitioning from a traditional HPC environment to a cloud-based service. This
transition is intended to enable the support of multiple simultaneous interactive simulations while
maintaining high-performance standards. The primary objective of this research is to broaden the
scope of applicable software within an HPC environment, ensuring that the transition to a cloud-
based service does not compromise performance efficacy. To achieve this goal, the study delves
into four distinct HPC load-balancing techniques. These techniques leverage virtualization,
software containers, and clustering to efficiently analyse, schedule, and execute game-based
simulation applications concurrently. The overarching aim is to determine the optimal approach
for extending HPC capabilities to accommodate the demands of multiple interactive simulations.
In the pursuit of the proposed HPC goal, the research places a particular emphasis on
experimenting with and evaluating these load-balancing techniques. Virtualization, software
containers, and clustering are assessed for their individual and collective performance in handling
the unique requirements of game-based simulations. The comparison of these techniques is critical
to understanding their respective strengths and weaknesses, thereby aiding in the determination of
the most suitable deployment technique. In making this determination, several factors come into
play. The availability of cluster resources, the number of competing software jobs, and the specific
characteristics of the software being scheduled all influence the choice of deployment technique.
By thoroughly investigating these considerations, the research aims to provide valuable insights
into the feasibility of extending HPC capabilities to a cloud-based service, ensuring that the chosen
approach aligns with the requirements of diverse simulation applications.
2.2 VIRTUAL MACHINE PLACEMENT IN CLOUD DATA CENTERS USING
A HYBRID MULTI-VERSE OPTIMIZATION ALGORITHM

Sasan Gharehpasha Cloud computing has revolutionized the landscape of computing by offering
a paradigm where a vast array of systems is interconnected in either private or public networks to
deliver dynamically scalable infrastructure for applications, data, and file storage. This
technological advancement has brought about a substantial reduction in the costs associated with
power consumption, application hosting, content storage, and resource delivery. Cloud computing
allows businesses to concentrate on their core goals without the need to continually expand
hardware resources, marking a significant shift in how computing resources are provisioned and
managed. One of the ongoing challenges in cloud computing, particularly in cloud data centers, is
the efficient placement of virtual machines on physical machines. The optimal placement of virtual
machines is crucial for managing resources effectively and preventing wastage. In this context, a
novel approach is introduced, combining the hybrid discrete multi-object whale optimization
algorithm with the multi-verse optimizer enhanced by chaotic functions. The primary objective of
this approach is twofold: firstly, to reduce power consumption by minimizing the number of active
physical machines in cloud data centers, and secondly, to enhance resource management by
strategically placing virtual machines over physical ones. By reducing power consumption and
preventing resource wastage through optimal virtual machine placement, this approach aims to
address critical issues in cloud data centre management. Moreover, it seeks to mitigate the
increasing rate of virtual migration to physical machines, contributing to the overall efficiency and
sustainability of cloud computing environments. To validate the efficacy of the proposed
algorithm, a comparative analysis is conducted against existing algorithms such as first fit,
VMPACS, and MBFD. The results obtained from this comparative study provide valuable insights
into the performance and superiority of the proposed approach in achieving optimal virtual
machine placement within cloud data centers. This research contributes to the ongoing efforts to
enhance the efficiency and sustainability of cloud computing infrastructures through innovative
algorithms and strategies.

2.3 AN ENERGY-AWARE HOST RESOURCE MANAGEMENT FRAMEWORK FOR

TWO-TIER VIRTUALIZED CLOUD DATA CENTERS
Chi Zhang The energy consumption of cloud data centers is a critical challenge that poses
constraints on the future growth of the cloud computing industry. This paper addresses this issue
with a specific focus on resource management within two-tier virtualized data centers, where
containers are deployed on virtual machines (VMs). The working model of the two-tier virtualized
data center is first defined, emphasizing the deployment of containers on VMs to enhance resource
utilization on hosts while ensuring the isolation and security of different jobs. To address the
energy consumption challenge, an energy-aware host resource management framework is
proposed. This framework comprises two key algorithms. The initial static placement algorithm
involves load balancing alternate placement and two-sided matching methods. These methods are
instrumental in efficiently placing containers onto VMs and VMs onto hosts, optimizing the
resource utilization of the entire system. The runtime dynamic consolidation algorithm, building
upon the initial placement, dynamically consolidates resources to utilize the least active hosts in
real-time, meeting the dynamic resource requirements of containers. Simulation experiments are
conducted using real workload traces to compare the proposed algorithms with existing ones. The
results demonstrate that the two algorithms exhibit superior performance in terms of host resource
utilization, the number of active hosts, the number of container migrations, and Service Level
Agreement (SLA) metrics. Importantly, the entire framework achieves a notable energy-saving
effect of at least 13.8%, showcasing its efficacy in addressing the energy consumption challenge
in cloud data centers. This research contributes to the broader discourse on sustainable and efficient
cloud computing by providing a detailed exploration of resource management strategies in two-
tier virtualized data centers. The demonstrated improvements in resource utilization and energy
efficiency underscore the potential of the proposed framework to contribute to the long-term
sustainability of cloud computing infrastructures.

2.4 AN EFFICIENT POWER-AWARE VM ALLOCATION MECHANISM IN CLOUD

DATA CENTERS: A MICRO GENETIC-BASED APPROACH

Mehran Trachoma In the ever-evolving landscape of cloud computing, optimizing power

efficiency in cloud servers has become a critical concern with far-reaching environmental
implications. The drive towards reducing greenhouse gas emissions has given rise to the concept
of green computing, where the focus is not only on enhancing computational performance but also
on minimizing the ecological footprint of data centre operations. A key strategy in achieving this
goal is the implementation of power-aware methods to strategically allocate virtual machines
(VMs) within the physical resources of data centers. Virtualization emerges as a promising
technology to facilitate power-aware VM allocation methods, given its ability to abstract and
manage resources efficiently. However, the allocation of VMs to physical hosts poses a complex
problem known to be NP-complete. In response to this challenge, evolutionary algorithms have
been employed as effective tools to tackle such optimization problems. This paper contributes to
the ongoing discourse by introducing a micro-genetic algorithm designed specifically for the
purpose of selecting optimal destinations among physical hosts for VM allocation. The micro-
genetic algorithm presented in this work is a sophisticated approach aimed at addressing the NP-
completeness of VM allocation. Through extensive evaluations in a simulation environment, the
study demonstrates the efficacy of the micro-genetic algorithm in significantly improving power
consumption metrics when compared to alternative methods. The results highlight the algorithm's
ability to make informed and efficient decisions in the allocation of VMs, leading to tangible
enhancements in power efficiency. By showcasing the valuable improvements achieved through
the micro-genetic algorithm, this research underscores the importance of leveraging evolutionary
approaches in solving complex optimization problems in the realm of cloud computing. The
findings contribute not only to the technical aspects of power-aware VM allocation but also to the
broader objective of establishing more sustainable and environmentally conscious practices within
data centre operations.

2.5 AN ENERGY AND SLA-AWARE RESOURCE MANAGEMENT STRATEGY IN

CLOUD DATA CENTERS

Chi Zhang The imperative for cloud providers to enhance investment yield by reducing the energy
consumption of data centers is balanced with the essential need to ensure that services delivered
meet diverse consumer requirements. This paper introduces a comprehensive resource
management strategy geared towards simultaneous reduction of energy consumption and
minimization of Service Level Agreement (SLA) violations in cloud data centers. The strategy
encompasses three refined methods tailored to address sub problems within the dynamic virtual
machine (VM) consolidation process. To enhance the effectiveness of host detection and improve
VM selection outcomes, the proposed strategy employs innovative approaches. Firstly, the
overloaded host detection method introduces a dynamic independent saturation threshold for each
host, accounting for CPU utilization trends. Secondly, the underutilized host detection method
integrates multiple factors beyond CPU utilization, incorporating the Naive Bayesian classifier to
calculate combined weights for hosts during prioritization. Lastly, the VM selection method takes
into account both current CPU usage and the anticipated future growth space of CPU demand for
VMs. The performance of the proposed strategy is rigorously evaluated through simulation in
CloudSim, and comparative analysis is conducted against five existing energy-saving strategies
using real-world workload traces. The experimental results demonstrate the superior performance
of the proposed strategy, showcasing minimal energy consumption and SLA violations in
comparison to other strategies. This outcome underscores the effectiveness of the refined methods
employed within the resource management strategy in achieving the dual objectives of energy
efficiency and SLA adherence. By surpassing existing energy-saving strategies in both energy
consumption reduction and SLA compliance, this research makes a significant contribution to the
ongoing efforts in optimizing resource management in cloud data centers. The proposed strategy
not only aligns with the imperative of energy efficiency but also underscores the importance of
delivering reliable services to consumers, thus striking a balance between environmental
sustainability and service quality in the realm of cloud computing.

2.6 RENEWABLE ENERGY-BASED MULTI-INDEXED JOB CLASSIFICATION AND

CONTAINER MANAGEMENT SCHEME FOR SUSTAINABILITY OF CLOUD DATA
CENTERS

Gagangeet Singh Aujla The landscape of modern computing has been significantly shaped by the
rise of Cloud Computing, offering on-demand services to end-users. However, the widespread use
of geo-distributed data centers to perform computing tasks raises concerns about the substantial
energy consumption associated with their operations. Addressing these energy-related challenges
in cloud environments has become imperative, and the integration of renewable energy resources,
coupled with strategic server selection and consolidation, presents a promising avenue for
mitigation. This paper introduces a novel approach, a renewable energy-aware multi-indexed job
classification and scheduling scheme, leveraging Container as-a-Service (CoaaS) for sustainability
in data centers. The proposed scheme aims to optimize energy usage by directing incoming
workloads from various devices to data centers equipped with a sufficient supply of renewable
energy. To achieve this, the paper outlines a renewable energy-based host selection and container
consolidation scheme, contributing to the overarching goal of energy efficiency and sustainability
in cloud computing environments. The effectiveness of the proposed scheme is rigorously
evaluated using real-world Google workload traces. The results demonstrate a substantial
improvement over existing schemes of similar categories, with energy savings reaching 15%, 28%,
and 10.55%. These findings underscore the viability and efficiency of the renewable energy-aware
multi-indexed job classification and scheduling scheme in enhancing sustainability within data
centers, while simultaneously achieving significant energy savings. By presenting a solution that
integrates renewable energy considerations, host selection, and container consolidation, this
research contributes to the ongoing discourse on sustainable practices in cloud computing. The
demonstrated improvements in energy savings validate the potential of the proposed scheme to not
only address current energy-related challenges but also to serve as a benchmark for future
advancements in environmentally conscious data centre operations.

2.7 A PLACEMENT ARCHITECTURE FOR A CONTAINER AS A SERVICE (CAAS) IN

A CLOUD ENVIRONMENT

Mohamed K The evolution of virtualization technologies has introduced containers as a

lightweight alternative to traditional virtual machines (VMs). Operating at the operating system
level, containers encapsulate tasks and their library dependencies for execution. The emerging
Container as a Service (CaaS) strategy is gaining prominence as a cloud service model. A critical
challenge within this paradigm is the placement of container instances on virtual machine
instances, constituting a classical scheduling problem. Previous research has often addressed either
virtual machine placement on physical machines (PMs), container placement, or task placement
without containerization in isolation. This approach, however, can lead to underutilized or over
utilized PMs and VMs. Consequently, there is a growing research interest in developing container
placement algorithms that consider the utilization of both instantiated VMs and used PMs
simultaneously. The primary objective of this study is to enhance resource utilization, focusing on
the number of CPU cores and memory size for both VMs and PMs, while minimizing the number
of instantiated VMs and active PMs in a cloud environment. The proposed placement architecture
employs scheduling heuristics, specifically Best Fit (BF) and Max Fit (MF), based on a fitness
function that concurrently evaluates the remaining resource waste of both PMs and VMs.
Additionally, a meta-heuristic placement algorithm is introduced, leveraging Ant Colony
Optimization based on Best Fit (ACO-BF) with the proposed fitness function. Experimental results
indicate that the proposed ACO-BF placement algorithm outperforms the BF and MF heuristics,
showcasing significant improvements in resource utilization for both VMs and PMs. By
incorporating a fitness function that considers the simultaneous evaluation of PMs and VMs, the
ACO-BF algorithm demonstrates its efficacy in optimizing resource allocation within a cloud
environment. This research contributes to the ongoing efforts to optimize container placement in
cloud environments, highlighting the importance of considering the utilization of both VMs and
PMs for enhanced resource efficiency. The proposed algorithms pave the way for more effective
and balanced resource utilization in cloud-based containerized applications.

2.8 ENERGY CONSUMPTION OPTIMIZATION OF CONTAINER-ORIENTED CLOUD

COMPUTING CENTER

Zhenjiang Li in the realm of container-based cloud computing, addressing the challenge of higher
energy consumption is crucial for optimizing overall efficiency. This paper introduces an enhanced
virtual migration strategy designed to specifically target and reduce energy consumption in
container-based cloud computing centers. The research methodology involves a comprehensive
examination of the intricate dependencies among physical machines, virtual machines, and
containers within a containerized environment of a cloud computing centre. The study proceeds
by analysing key factors that exert influence on the energy consumption of the data canter, taking
into account the dependencies identified in the container environment. Subsequently, a
mathematical model is established to encapsulate these dependencies and facilitate a systematic
understanding of the energy consumption dynamics. Building upon this analysis, the paper
proposes an optimal utilization priority algorithm. This algorithm is designed based on the non-
linear relationship between utilization and energy consumption across different physical machines,
aiming to prioritize and allocate resources in a manner that minimizes energy consumption.
Simulation experiments conducted on the Container CloudSim platform substantiate the
effectiveness of the proposed method. The results demonstrate that the introduced approach
significantly reduces the energy consumption of the data centre compared to traditional strategies
such as random scheduling and maximum utilization. The emphasis on practical differences
between physical machines in real-world environments sets this improved virtual migration
strategy apart, offering a tailored solution to address the energy consumption challenges in
container-based cloud computing. In conclusion, this paper not only refines classical virtual
migration strategies but also introduces a pragmatic and easily implementable algorithm for
reducing energy consumption in containerized cloud computing environments. The simplicity and
efficacy of the proposed approach make it a noteworthy candidate for widespread adoption,
contributing to the ongoing efforts to make container-based cloud computing more energy-
efficient and sustainable.

2.9 AN ENERGY, PERFORMANCE EFFICIENT RESOURCE CONSOLIDATION

SCHEME FOR HETEROGENEOUS CLOUD DATACENTERS

Ayaz Ali Khan Datacentres, as the primary electricity consumers for cloud computing, play a
pivotal role in providing the IT backbone for contemporary businesses and economies. However,
studies indicate that a significant portion of servers in U.S. datacentres are underutilized or idle,
presenting an opportunity for energy savings through resource consolidation techniques. The
challenge lies in the fact that consolidation, involving migrations of virtual machines (VMs),
containers, and/or applications, can be costly in terms of both energy consumption and
performance loss. This paper addresses this challenge by proposing a consolidation algorithm that
prioritizes the most effective migration among VMs, containers, and applications. Additionally,
the study investigates how migration decisions can be made to save energy without negatively
impacting service performance. Through a series of experiments utilizing real workload traces for
800 hosts, approximately 1,516 VMs, and over a million containers, the paper evaluates the impact
of different migration approaches on datacentre energy consumption and performance. The
findings highlight a trade-off between migrating containers and virtual machines, where migrating
virtual machines tends to be more performance-efficient, while migrating containers can be more
energy-efficient. The study also suggests that migrating containerized applications, running within
virtual machines, could lead to an energy and performance-efficient consolidation technique in
large-scale datacentres. The evaluation results indicate that migrating applications may be
approximately 5.5% more energy-efficient and 11.9% more performance-efficient than VM
migration. Furthermore, energy and performance-efficient consolidation is approximately 14.6%
more energy-efficient and 7.9% more performance-efficient than application migration. The study
generalizes these findings through repeatable experiments across various workloads, resources,
and datacentre setups. In conclusion, the research sheds light on the nuanced trade-offs involved
in migration decisions for datacentre consolidation. By proposing a consolidation algorithm and
providing insights into the energy and performance efficiencies of different migration approaches,
the paper contributes to the ongoing efforts to optimize resource usage in large-scale datacentres.

2.10 HEPORCLOUD: AN ENERGY AND PERFORMANCE EFFICIENT RESOURCE

ORCHESTRATOR FOR HYBRID HETEROGENEOUS CLOUD COMPUTING
ENVIRONMENTS

Ayaz Ali Khan in major Information Technology (IT) companies like Google, Rackspace, and
Amazon Web Services (AWS), the execution of customers' workloads and applications relies
heavily on virtualization and containerization technologies. These companies operate large-scale
datacentres that provide computational resources, but the substantial energy consumption of these
datacentres raises ecological concerns. Each company employs different approaches, with Google
utilizing containers, Rackspace offering bare-metal hardware, and AWS employing a mix of
virtual machines (VMs), containers (ECS), and containers inside VMs (Lambda). This diversity in
technology usage makes resource management a complex task. Effective resource management is
crucial, especially in hybrid platforms where various sandboxing technologies like bare-metal,
VMs, containers, and nested containers coexist. The absence of centralized, workload-aware
resource managers and consolidation policies raises questions about datacentres energy efficiency,
workload performance, and user costs. This paper addresses these concerns through a series of
experiments using Google workload data for 12,583 hosts and approximately one million tasks
across four different types of workloads. The focus is on demonstrating the potential benefits of
using workload-aware resource managers in hybrid clouds, achieving energy and cost savings in
heterogeneous hybrid datacentres without negatively impacting workload performance. The paper
also explores how different allocation policies, combined with various migration approaches,
impact datacentres energy and performance efficiencies. The empirical evaluation, based on
plausible assumptions for hybrid datacentres setups, reveals compelling results. In scenarios with
no migration, a single scheduler is found to be up to 16.86% more energy-efficient than distributed
schedulers. However, when migrations are considered, the proposed resource manager
demonstrates the potential to save up to 45.61% energy and improve workload performance by up
to 17.9%. In conclusion, the research highlights the significance of workload-aware resource
managers in optimizing energy efficiency and cost savings in heterogeneous hybrid datacentres.
The findings provide valuable insights for IT companies seeking to enhance the performance and
sustainability of their datacentres operations.
 CHAPTER 3

SYSTEM ANALYSIS

3.1 EXISTING SYSTEM

In distributed environments, cloud computing is widely used to manage user requests for resources
and services. Resource scheduling is used to handle user requests for resources based on priorities
within a given time frame. In today’s environment, every industry management rely on smart
devices connected to the internet. These devices deal with the massive amounts of data processed
and detected by smart medical sensors without sacrificing performance factors like throughput and
latency. This has prompted the requirement for load balancing among the smart operational
devices to prevent any insensitivity. Load balancing is used to manage large amounts of data in
both a centralized and distributed manner. We use reinforcement learning algorithms such as GA,
SARSA, and Q-learning for resource scheduling. These algorithms are used to predict the optimal
solution to manage load in cloud-based applications . The main drawbacks in existing system is
less security and low performance

3.1.1 DRAWBACKS

1. Existing systems exhibit vulnerabilities leading to potential data breaches and privacy
concerns.

2. Systems experience suboptimal performance due to inefficient resource allocation and

management.

3. Limited scalability hampers the ability to handle increasing data volumes and user
demands effectively.

4. System reliability is compromised, leading to potential downtimes and disruptions in

cloud applications.
3.2 PROPOSED SYSTEM

The proposed system introduces an innovative approach to Cloud VM scheduling by addressing

the limitations of current instant-based resource allocation. Leveraging historical data on VM
resource utilization, the system employs a scheduling algorithm powered by Particle Swarm
Optimization (PSO). This methodology enables the system to learn and adapt to the dynamic
behavior of the Cloud environment over time. Unlike traditional approaches, the proposed system
prioritizes overall and long-term resource utilization, aiming to minimize the impact of Cloud
management processes on deployed VMs. By optimizing performance and reducing the count of
physical machines through the PSO classifier, the system achieves enhanced efficiency and
maximizes real CPU utilization, thereby refining the conventional VM placement strategies in
Cloud systems.

3.2.1 ADVANTAGES
• The algorithm takes into account the already running VM resource usage over time to
optimize the placement of VMs. This can lead to improved performance for the VMs, as
they are less likely to be placed on hosts that are already overloaded.

• The algorithm uses PSO, which is a metaheuristic algorithm that is known for its ability to
find good solutions to complex problems. This makes the algorithm more likely to find a
good VM placement solution, even in large and complex cloud systems.

• It is computationally efficient, which makes it suitable for large-scale cloud systems.

• It can be configured to meet a variety of objectives, such as minimizing costs, maximizing

performance, and minimizing the number of VM migrations.

3.3 FEASIBILITY STUDY

Preliminary investigation examines project feasibility; the likelihood the system will be useful to
the organization. The main objective of the feasibility study is to test the Technical, Operational
and Economical feasibility for adding new modules and debugging old running system. All system
is feasible if they are unlimited resources and infinite time. There are aspects in the feasibility
study portion of the preliminary investigation:

 Technical Feasibility
 Operation Feasibility
 Economical Feasibility

3.3.1 TECHNICAL FEASIBILITY

The technical issue usually raised during the feasibility stage of the investigation includes
the following:

 Does the necessary technology exist to do what is suggested?

 Do the proposed equipments have the technical capacity to hold the data required to use the
new system?
 Will the proposed system provide adequate response to inquiries, regardless of the number or
location of users?
 Can the system be upgraded if developed?
 Are there technical guarantees of accuracy, reliability, ease of access and data security?
Earlier no system existed to cater to the needs of ‘Secure Infrastructure Implementation
System’. The current system developed is technically feasible. It is a web based user interface for
audit workflow at DB2 Database. Thus it provides an easy access to the users. The database’s
purpose is to create, establish and maintain a workflow among various entities in order to facilitate
all concerned users in their various capacities or roles. Permission to the users would be granted
based on the roles specified.

Therefore, it provides the technical guarantee of accuracy, reliability and security. The
software and hard requirements for the development of this project are not many and are already
available in-house at NIC or are available as free as open source. The work for the project is done
with the current equipment and existing software technology. Necessary bandwidth exists for
providing a fast feedback to the users irrespective of the number of users using the system.
3.3.2 OPERATIONAL FEASIBILITY

Proposed projects are beneficial only if they can be turned out into information system.
That will meet the organization’s operating requirements. Operational feasibility aspects of the
project are to be taken as an important part of the project implementation. Some of the important
issues raised are to test the operational feasibility of a project includes the following: -

 Is there sufficient support for the management from the users?

 Will the system be used and work properly if it is being developed and implemented?
 Will there be any resistance from the user that will undermine the possible application benefits?
This system is targeted to be in accordance with the above-mentioned issues. Beforehand,
the management issues and user requirements have been taken into consideration. So there is no
question of resistance from the users that can undermine the possible application benefits.

The well-planned design would ensure the optimal utilization of the computer resources and would
help in the improvement of performance status.

3.3.3 ECONOMIC FEASIBILITY

A system can be developed technically and that will be used if installed must still be a good
investment for the organization. In the economic feasibility, the development cost in creating the
system is evaluated against the ultimate benefit derived from the new systems. Financial benefits
must equal or exceed the costs.

The system is economically feasible. It does not require any addition hardware or software.
Since the interface for this system is developed using the existing resources and technologies
available at NIC, there is nominal expenditure and economical feasibility for certain.
 CHAPTER 4

SYSTEM SPECIFICATION

4.1 HARDWARE REQUIREMENTS

CPU type : Intel core i5 processor

Clock speed : 3.0 GHz

RAM size : 8 GB

Hard disk capacity : 500 GB

Keyboard type : Internet Keyboard

CD -drive type : 52xmax

4.2 SOFTWARE REQUIREMENTS

Operating System : Windows 10

Front End : JAVA

 CHAPTER 5

SOFTWARE DESCRIPTION

5.1 FRONT END: JAVA

The software requirement specification is created at the end of the analysis task. The function and
performance allocated to software as part of system engineering are developed by establishing a
complete information report as functional representation, a representation of system behavior, an
indication of performance requirements and design constraints, appropriate validation criteria.

FEATURES OF JAVA

Java platform has two components:

 The Java Virtual Machine (Java VM)

 The Java Application Programming Interface (Java API)
The Java API is a large collection of ready-made software components that provide many useful
capabilities, such as graphical user interface (GUI) widgets. The Java API is grouped into libraries
(packages) of related components.

The following figure depicts a Java program, such as an application or applet, that's running on
the Java platform. As the figure shows, the Java API and Virtual Machine insulates the Java
program from hardware dependencies.

As a platform-independent environment, Java can be a bit slower than native code.

However, smart compilers, well-tuned interpreters, and just-in-time byte code compilers can bring
Java's performance close to that of native code without threatening portability.
SOCKET OVERVIEW:

A network socket is a lot like an electrical socket. Various plugs around the network have
a standard way of delivering their payload. Anything that understands the standard protocol can
“plug in” to the socket and communicate.

Internet protocol (IP) is a low-level routing protocol that breaks data into small packets and
sends them to an address across a network, which does not guarantee to deliver said packets to the
destination.

Transmission Control Protocol (TCP) is a higher-level protocol that manages to reliably transmit
data. A third protocol, User DatagramProtocol (UDP), sits next to TCP and can be used directly to
support fast, connectionless, unreliable transport of packets.

CLIENT/SERVER:

A server is anything that has some resource that can be shared. There are compute
servers, which provide computing power; print servers, which manage a collection of printers; disk
servers, which provide networked disk space; and web servers, which store web pages. A client is
simply any other entity that wants to gain access to a particular server.

A server process is said to “listen” to a port until a client connects to it. A server is
allowed to accept multiple clients connected to the same port number, although each session is
unique. To manage multiple client connections, a server process must be multithreaded or have
some other means of multiplexing the simultaneous I/O.

RESERVED SOCKETS:

Once connected, a higher-level protocol ensues, which is dependent on which port

user are using. TCP/IP reserves the lower, 1,024 ports for specific protocols. Port number 21 is for
FTP, 23 is for Telnet, 25 is for e-mail, 79 is for finger, 80 is for HTTP, 119 is for Netnews-and the
list goes on. It is up to each protocol to determine how a client should interact with the port.

JAVA AND THE NET:

 Java supports TCP/IP both by extending the already established stream I/O
interface. Java supports both the TCP and UDP protocol families. TCP is used for reliable stream-
based I/O across the network. UDP supports a simpler, hence faster, point-to-point datagram-
oriented model.

INETADDRESS:

The InetAddress class is used to encapsulate both the numerical IP address and the
domain name for that address. User interacts with this class by using the name of an IP host, which
is more convenient and understandable than its IP address. The InetAddress class hides the number
inside. As of Java 2, version 1.4, InetAddress can handle both IPv4 and IPv6 addresses.

FACTORY METHODS:

The InetAddress class has no visible constructors. To create an InetAddress object,

user use one of the available factory methods. Factory methods are merely a convention whereby
static methods in a class return an instance of that class. This is done in lieu of overloading a
constructor with various parameter lists when having unique method names makes the results
much clearer.

Three commonly used InetAddress factory methods are:

1. Static InetAddressgetLocalHost ( ) throws

UnknownHostException

2. Static InetAddressgetByName (String hostName)

throwsUnknowsHostException

3. Static InetAddress [ ] getAllByName (String hostName)

throwsUnknownHostException

INSTANCE METHODS:

The InetAddress class also has several other methods, which can be used on the
objects returned by the methods just discussed. Here are some of the most commonly used.
 Boolean equals (Object other)- Returns true if this object has the same
Internet address as other.

1. byte [ ] get Address ( )- Returns a byte array that represents the object’s
Internet address in network byte order.

2. String getHostAddress ( ) - Returns a string that represents the host address

associated with the InetAddress object.

3. String get Hostname ( ) - Returns a string that represents the host name associated with
the InetAddress object.

4. booleanisMulticastAddress ( )- Returns true if this Internet address is a multicast

address. Otherwise, it returns false.

5. String toString ( ) - Returns a string that lists the host name and the IP address for
convenience.

TCP/IP CLIENT SOCKETS:

TCP/IP sockets are used to implement reliable, bidirectional, persistent, point-

to-point and stream-based connections between hosts on the Internet. A socket can be used to
connect Java’s I/O system to other programs that may reside either on the local machine or on any
other machine on the Internet.

There are two kinds of TCP sockets in Java. One is for servers, and the other
is for clients. The Server Socket class is designed to be a “listener,” which waits for clients to
connect before doing anything. The Socket class is designed to connect to server sockets and
initiate protocol exchanges.

The creation of a Socket object implicitly establishes a connection between the client and
server. There are no methods or constructors that explicitly expose the details of establishing that
connection. Here are two constructors used to create client sockets
Socket (String hostName, intport) - Creates a socket connecting the local host to the named host
and port; can throw an UnknownHostException or anIOException.

Socket (InetAddressipAddress, intport) - Creates a socket using a

preexistingInetAddressobject and a port; can throw an IOException.

A socket can be examined at any time for the address and port information
associated with it, by use of the following methods:

 InetAddressgetInetAddress ( ) - Returns the InetAddress associated with the

Socket object.
 IntgetPort ( ) - Returns the remote port to which this Socket object is
connected.
 IntgetLocalPort ( ) - Returns the local port to which this Socket object is
connected.
Once the Socket object has been created, it can also be examined to gain access to
the input and output streams associated with it. Each of these methods can throw an IO Exception
if the sockets have been invalidated by a loss of connection on the Net.

Input Streamget Input Stream ( ) - Returns the InputStream associated with the
invoking socket.

Output Streamget Output Stream ( ) - Returns the OutputStream associated with the
invoking socket.

TCP/IP SERVER SOCKETS:

Java has a different socket class that must be used for creating server applications. The
ServerSocket class is used to create servers that listen for either local or remote client programs to
connect to them on published ports. ServerSockets are quite different form normal Sockets.

When the user create a ServerSocket, it will register itself with the system as having an interest in
client connections.
  ServerSocket(int port) - Creates server socket on the specified port with a queue length of
50.
 Serversocket(int port, int maxQueue) - Creates a server socket on the specified portwith a
maximum queue length of maxQueue.
 ServerSocket(int port, int maxQueue, InetAddress localAddress)-Creates a server socket
on the specified port with a maximum queue length of maxQueue. On a multihomed host,
localAddress specifies the IP address to which this socket binds.
 ServerSocket has a method called accept( ) - which is a blocking call that will wait for a
client to initiate communications, and then return with a normal Socket that is then used
for communication with the client.
URL:

The Web is a loose collection of higher-level protocols and file formats, all unified in a
web browser. One of the most important aspects of the Web is that Tim Berners-Lee devised a
saleable way to locate all of the resources of the Net. The Uniform Resource Locator (URL) is
used to name anything and everything reliably.

The URLprovides a reasonably intelligible form to uniquely identify or address

information on the Internet. URLs are ubiquitous; every browser uses them to identify information
on the Web.
 CHAPTER 6

PROJECT DESCRIPTION

6.1 PROBLEM DEFINITION

In contemporary cloud-based applications, the integration of smart devices and sensors has led to
a surge in data volumes, necessitating efficient resource management strategies. However, existing
systems suffer from critical drawbacks, notably in security and performance. Security
vulnerabilities pose significant risks, potentially leading to data breaches and compromising
patient privacy. Additionally, suboptimal performance due to inefficient resource allocation
undermines system responsiveness and user satisfaction. Addressing these challenges is
imperative to ensure the reliability, scalability, and security of cloud-based healthcare systems,
facilitating effective management of user requests and data while upholding the highest standards
of privacy and performance.

6.2 MODULE DESCRIPTION

6.2.1 VM SCHEDULING

VM scheduling is a pivotal module within cloud computing systems that orchestrates the allocation
of virtual machines (VMs) to physical hosts. The primary objective of VM scheduling is to
optimize resource utilization and enhance overall system performance. This module considers
instantaneous resource usage, historical utilization patterns, and long-term performance metrics to
make informed decisions about VM placement. The effectiveness of VM scheduling directly
impacts the efficiency of cloud environments by ensuring that computational resources are
allocated judiciously, adapting dynamically to varying workloads.

6.2.2 DATA ANALYSIS

Data analysis is a crucial role in the proposed system, providing the foundation for informed
decision-making. This module involves the examination and interpretation of historical VM
resource utilization data over time. By employing statistical and machine learning techniques, data
analysis contributes to understanding the patterns and trends within the cloud environment.
Insights derived from this analysis inform the VM scheduling algorithm, allowing it to adapt to
changing conditions and optimize resource allocation based on past performance metrics.

6.2.3 CLASSIFICATION ALGORITHM

The classification algorithm is a key component employed to categorize and organize data in the
context of VM scheduling. This module likely involves the application of machine learning
techniques to classify VMs based on their resource utilization characteristics. The algorithm's
ability to distinguish between different classes of VMs is crucial for making informed decisions
about their placement and resource allocation within the cloud infrastructure.

6.2.4 PARTICLE SWARM OPTIMIZATION (PSO)

Particle Swarm Optimization (PSO) is a sophisticated optimization algorithm utilized in the

proposed system. This module involves the application of PSO to fine-tune the VM scheduling
process. PSO simulates the social behavior of particles in a swarm, each representing a potential
solution. By leveraging the collective intelligence of the swarm, PSO dynamically adapts the VM
scheduling algorithm, optimizing parameters to achieve enhanced efficiency and performance.

6.2.5 OPTIMIZATION SCHEME

The optimization scheme serves as the overarching framework that integrates the various
components of the proposed system. This module encapsulates the strategy for enhancing system
performance, which may include the coordination of VM scheduling, data analysis, classification
algorithms, and PSO. The optimization scheme aims to minimize the impact of management
processes on deployed VMs by maximizing real CPU utilization and reducing the count of physical
machines. It provides a holistic approach to refining VM placement strategies and ensuring
efficient resource utilization in cloud computing environments.
6.3 SYSTEM FLOW DIAGRAM

VM OPTIMIZATION VM SELECTION
FEATURES
SCHDULING SCHEMES PHASE
OPTIMIZATION
SCHEME

CLASSIFICATIO LABEL
PSO
N ALGORITHM

VM resource monitoring
process VM scheduling

MONITORING WORK LOAD

6.4 INPUT DESIGN

The proposed energy-aware host resource management framework for virtual machines in cloud
data centers using the particle swarm optimization (PSO) algorithm requires the following inputs:

1. Resource usage information: This information includes the CPU, memory, and network
usage of the VMs and hosts. This information is collected by the resource monitor
component of the framework.

2. VM requirements: This information includes the CPU, memory, and network requirements
of the VMs. This information is typically provided by the cloud provider or the user.

3. Host configurations: This information includes the CPU, memory, and network capacities
of the hosts. This information is typically stored in the data centre’s management system.

4. Energy consumption model: This model is used to estimate the energy consumption of the
hosts based on their resource usage. This model can be based on historical data or on a
theoretical model.

6.5 OUTPUT DESIGN

The proposed energy-aware host resource management framework for virtual machines in cloud
data centers using the particle swarm optimization (PSO) algorithm produces the following
outputs:

1. VM placement: This is the mapping of VMs to hosts. The framework determines the
optimal placement of VMs based on their resource requirements, the available resources
of the hosts, and the energy consumption of the hosts.

2. VM scheduling: This is the scheduling of VMs on hosts. The framework determines the
optimal scheduling of VMs on hosts based on their resource requirements, the available
resources of the hosts, and the energy consumption of the hosts.

3. Energy consumption estimates: This is an estimate of the energy consumption of the data
centre based on the VM placement and scheduling. The framework uses an energy
 consumption model to estimate the energy consumption of the hosts based on their resource
usage.

4. Performance metrics: This is a set of metrics that measure the performance of the data
centre, such as throughput, latency, and response time. The framework can be used to
monitor the performance of the data centre and to identify any potential problems.
 CHAPTER 7

7. SYSTEM TESTING AND IMPLEMENTATION

7.1 SYSTEM TESTING

To verify that the proposed framework meets the following requirements:

 Functionality: The framework should be able to correctly place and schedule VMs on
hosts.

 Performance: The framework should be able to place and schedule VMs in a timely
manner.

 Energy Efficiency: The framework should be able to reduce energy consumption compared
to traditional methods.

 Scalability: The framework should be able to scale to large cloud data centers.

 Reliability: The framework should be able to handle failures and recover gracefully.

7.2 SYSTEM IMPLEMENTATION

1. Resource Monitor: The resource monitor collects resource usage information from the
VMs and hosts. This information includes CPU, memory, and network usage. The resource
monitor can be implemented using a variety of tools, such as SNMP or IPMI.

2. Decision Maker: The decision maker uses the PSO algorithm to determine the optimal VM
placement and scheduling. The decision maker can be implemented using a variety of
programming languages, such as Python or Java.

3. Executor: The executor enforces the decision maker's decisions. The executor can be
implemented using a variety of tools, such as OpenStack or Cloud Stack.
 CHAPTER 8

SYSTEM MAINTENANCE

The objectives of this maintenance work are to make sure that the system gets into work all time
without any bug. Provision must be for environmental changes which may affect the computer or
software system. This is called the maintenance of the system. Nowadays there is the rapid change
in the software world. Due to this rapid change, the system should be capable of adapting these
changes. In this project the process can be added without affecting other parts of the system.
Maintenance plays a vital role. The system is liable to accept any modification after its
implementation. This system has been designed to favor all new changes. Doing this will not affect
the system’s performance or its accuracy.

Maintenance is necessary to eliminate errors in the system during its working life and to tune the
system to any variations in its working environment. It has been seen that there are always some
errors found in the system that must be noted and corrected. It also means the review of the system
from time to time.

The review of the system is done for:

 Knowing the full capabilities of the system.

 Knowing the required changes or the additional requirements.

 Studying the performance.

TYPES OF MAINTENANCE:

 Corrective maintenance

 Adaptive maintenance

 Perfective maintenance

 Preventive maintenance
8.1 CORRECTIVE MAINTENANCE

Changes made to a system to repair flows in its design coding or implementation. The design
of the software will be changed. The corrective maintenance is applied to correct the errors that
occur during that operation time. The user may enter invalid file type while submitting the
information in the particular field, then the corrective maintenance will displays the error message
to the user in order to rectify the error.

Maintenance is a major income source. Nevertheless, even today many organizations assign
maintenance to unsupervised beginners, and less competent programmers.

The user’s problems are often caused by the individuals who developed the product, not
the maintainer. The code itself may be badly written maintenance is despised by many software
developers unless good maintenance service is provided, the client will take future development
business elsewhere. Maintenance is the most important phase of software production, the most
difficult and most thankless.

8.2 ADAPTIVE MAINTENANCE:

It means changes made to system to evolve its functionalities to change business

needs or technologies. If any modification in the modules the software will adopt those
modifications. If the user changes the server then the project will adapt those changes. The
modification server work as the existing is performed.

8.3 PERFECTIVE MAINTENANCE:

Perfective maintenance means made to a system to add new features or improve

performance. The perfective maintenance is done to take some perfect measures to maintain the
special features. It means enhancing the performance or modifying the programs to respond to the
users need or changing needs. This proposed system could be added with additional functionalities
easily. In this project, if the user wants to improve the performance further then this software can
be easily upgraded.

8.4 PREVENTIVE MAINTENANCE:

Preventive maintenance involves changes made to a system to reduce the changes of features
system failure. The possible occurrence of error that might occur are forecasted and prevented with
suitable preventive problems. If the user wants to improve the performance of any process then
the new features can be added to the system for this project.
 CHAPTER 9

9. CONCLUSION

In conclusion, the developed Cloud VM scheduling algorithm, utilizing historical VM resource

utilization data and Particle Swarm Optimization (PSO), presents a promising solution to the
challenges associated with instant-based resource allocation in cloud systems. The system's ability
to learn and adapt to the evolving behavior of the environment over time contributes to improved
efficiency and performance optimization. By prioritizing overall and long-term resource utilization
and minimizing the impact of management processes on deployed VMs, the proposed approach
offers a refined strategy for VM placement. The demonstrated reduction in the count of physical
machines underscores the system's effectiveness in resource allocation.

FUTURE WORK

For future work, the proposed Cloud VM scheduling algorithm lays the foundation for several
potential enhancements and research directions. Further exploration could involve the integration
of machine learning techniques to continuously adapt the scheduling algorithm based on real-time
changes in the Cloud environment. Additionally, the system could benefit from considering energy
efficiency aspects to align with the growing importance of sustainable computing. Exploring the
application of the proposed algorithm in diverse Cloud architectures and scaling it for larger and
more complex environments would provide insights into its scalability and generalizability.
 CHAPTER 10

APPENDICES

10.1 SOURCE CODE

package power;

import java.util.List;

import java.util.ArrayList;

import org.cloudbus.cloudsim.Vm;

import org.cloudbus.cloudsim.Host;

import org.cloudbus.cloudsim.power.PowerHost;

import org.cloudbus.cloudsim.power.PowerVm;

/**

* @author admin

*/

public class Details

static String vms[][];

static String host[][];

static ArrayList Vt=new ArrayList();

 static ArrayList Ht=new ArrayList();

//static List<Vm> vmlist=new ArrayList<Vm>();

static List<PowerVm> vmlist=new ArrayList<PowerVm>();

//static List<Host> hostList=new ArrayList<Host>(); //

static List<PowerHost> hostList=new ArrayList<PowerHost>(); // PM

static double Velocity[][];

static double Position[][];

static ArrayList request=new ArrayList();

static ArrayList population=new ArrayList();

static ArrayList initialpop=new ArrayList();

static int pop=8;

static double pbest[]=new double[pop];

static double gbest=200;

static String psobest;

static List<PowerVm> allVM=new ArrayList<PowerVm>();

static ArrayList newList=new ArrayList();

}
package power;

import java.awt.Color;

import org.jfree.chart.ChartFactory;

import org.jfree.chart.ChartFrame;

import org.jfree.chart.JFreeChart;

import org.jfree.chart.plot.CategoryPlot;

import org.jfree.chart.plot.PlotOrientation;

import org.jfree.chart.renderer.category.CategoryItemRenderer;

import org.jfree.data.category.DefaultCategoryDataset;

/**

* @author admin

*/

public class Graph1

public void display1(double val)

try

DefaultCategoryDataset dataset = new DefaultCategoryDataset();

dataset.setValue(183, "Existing" ,"Execution Time");

dataset.setValue(val, "Proposed" ,"Execution Time");

JFreeChart chart = ChartFactory.createBarChart

("Execution Time","", "Time in ms", dataset,

PlotOrientation.VERTICAL, true,true, false);

chart.getTitle().setPaint(Color.blue);

CategoryPlot p = chart.getCategoryPlot();

p.setRangeGridlinePaint(Color.red);

System.out.println("Range : "+p.getRangeAxisCount() );

CategoryItemRenderer renderer = p.getRenderer();

renderer.setSeriesPaint(0, Color.red);

renderer.setSeriesPaint(1, Color.green);

// renderer.setSeriesPaint(3, Color.yellow);
 ChartFrame frame1=new ChartFrame("Execution Time",chart);

frame1.setSize(400,400);

frame1.setVisible(true);

catch(Exception e)

e.printStackTrace();

public void display2(double val)

try

DefaultCategoryDataset dataset = new DefaultCategoryDataset();

dataset.setValue(2.7401, "Existing" ,"Energy Consumption");

dataset.setValue(val, "Proposed" ,"Energy Consumption");

JFreeChart chart = ChartFactory.createBarChart

("Energy Consumption","", "Value", dataset,

PlotOrientation.VERTICAL, true,true, false);

chart.getTitle().setPaint(Color.blue);

CategoryPlot p = chart.getCategoryPlot();

p.setRangeGridlinePaint(Color.red);

System.out.println("Range : "+p.getRangeAxisCount() );

CategoryItemRenderer renderer = p.getRenderer();

renderer.setSeriesPaint(0, Color.BLUE);

renderer.setSeriesPaint(1, Color.pink);

// renderer.setSeriesPaint(3, Color.yellow);

ChartFrame frame1=new ChartFrame("Energy Consumption",chart);

 frame1.setSize(400,400);

frame1.setVisible(true);

catch(Exception e)

e.printStackTrace();

package power;

/**

* @author admin

*/

public class Main {

/**

* @param args the command line arguments

*/

public static void main(String[] args) {

// TODO code application logic here

long tm1=System.currentTimeMillis();
 VMAllocation vm=new VMAllocation();

vm.readVM();

vm.readHost();

vm.createHost();

vm.createVM();

vm.optimiseVmAllocation();

long tm2=System.currentTimeMillis();

long tim=tm2-tm1;

System.out.println(tim);

Graph1 gr=new Graph1();

gr.display1(tim);

gr.display2(1.4332);

package power;

import java.util.Random;

import org.cloudbus.cloudsim.power.PowerHost;

import org.cloudbus.cloudsim.power.PowerVm;

import org.cloudbus.cloudsim.Vm;

import org.cloudbus.cloudsim.power.PowerVmAllocationPolicySimple;

import java.util.List;

import java.util.Map;

/**

*
* @author admin

*/

public class PSO

Details dt=new Details();

double weight=0.1;

double c1=1;

double c2=1;

double x_max = 4.0;

double v_max=4.0;

double x_min = -0.4;

double v_min=-4.0;

int iter=50;

PSO()

public void applyPSO()

try

Random rn=new Random();

dt.Velocity=new double[dt.pop][dt.request.size()];

dt.Position=new double[dt.pop][dt.request.size()];

for(int it=0;it<iter;it++)

int rk[][]=new int[dt.pop][dt.request.size()];

for(int i=0;i<dt.request.size();i++)

dt.Position[0][i]=x_min + (x_max - x_min) *rn.nextDouble() ;

dt.Velocity[0][i]=v_min + (v_max - v_min) *rn.nextDouble() ;

double de[][]=new double[dt.request.size()][3];

for(int i=0;i<dt.request.size();i++)

de[i][0]=dt.Position[0][i];

de[i][1]=i;

de[i][2]=i;

for(int i=0;i<dt.request.size();i++)

for(int j=i+1;j<dt.request.size();j++)

{
 if(de[i][0]>de[j][0])

double t1=de[i][0];

de[i][0]=de[j][0];

de[j][0]=t1;

double t2=de[i][1];

de[i][1]=de[j][1];

de[j][1]=t2;

for(int i=0;i<dt.request.size();i++)

int k1=(int)de[i][1];

int k2=(int)de[i][2];

rk[0][k1]=(k2%dt.host.length)+1;

for(int pi=0;pi<dt.pop;pi++)

String g1[]=dt.population.get(pi).toString().split("#");

double Cexe=0;
 for(int i=0;i<g1.length;i++)

String g2[]=dt.request.get(i).toString().split("#");

double dur=Double.parseDouble(g2[1]);

double

res=Double.parseDouble(g2[2])+Double.parseDouble(g2[3])+Double.parseDouble(g2[4]);

if(res==0)

res=1;

Cexe=Cexe+(Double.parseDouble(g1[i])*(dur/res));

Cexe=Cexe+(dt.Position[pi][i]-dt.Velocity[pi][i]);

// System.out.println("pp= "+Cexe);

dt.pbest[pi]=Cexe;

if(dt.gbest<Cexe)

dt.psobest=dt.population.get(pi).toString();

dt.gbest=Cexe;

for(int i=0;i<dt.pop-1;i++)

for(int j=0;j<dt.request.size();j++)

{
 dt.Velocity[i+1][j]=weight*dt.Velocity[i][j]+c1*rn.nextDouble()*(dt.pbest[i]-

dt.Position[i][j])+c2*rn.nextDouble()*(dt.gbest-dt.Position[i][j]);

dt.Position[i+1][j]=dt.Position[i][j]+dt.Velocity[i][j];

} // iter

System.out.println("gbest final "+dt.gbest);

System.out.println("pso best "+dt.psobest);

catch(Exception e)

e.printStackTrace();

public double fittnessFun(PowerHost ph)

double uti=0;

try

PowerVmAllocationPolicySimple ps=new
PowerVmAllocationPolicySimple(dt.hostList);

int h=0;

for(int i=0;i<dt.vmlist.size();i++)

PowerVm vm=dt.vmlist.get(i);

if(!dt.allVM.contains(vm))

boolean bool=ps.allocateHostForVm(vm, ph);

if(bool)

uti=ph.getUtilizationOfRam()+ph.getUtilizationOfBw()+ph.getUtilizationOfCpuMips();

dt.allVM.add(vm);

dt.newList.add(vm.getId()+"#"+ph.getId());

h++;

else

System.out.println("VM - "+vm.getId()+" is migrated");

break;

}
 }

//System.out.println(ph.getId()+" : "+uti);

catch(Exception e)

e.printStackTrace();

return uti;

package power;

import java.io.File;

import java.io.FileInputStream;

import java.util.ArrayList;

import java.util.Calendar;

import java.util.LinkedList;

import java.util.List;

import org.cloudbus.cloudsim.CloudletSchedulerTimeShared;

import org.cloudbus.cloudsim.CloudletScheduler;

import org.cloudbus.cloudsim.Datacenter;

import org.cloudbus.cloudsim.DatacenterCharacteristics;

import org.cloudbus.cloudsim.Log;

import org.cloudbus.cloudsim.Pe;

import org.cloudbus.cloudsim.Storage;
import org.cloudbus.cloudsim.Vm;

import org.cloudbus.cloudsim.VmAllocationPolicySimple;

import org.cloudbus.cloudsim.VmSchedulerTimeShared;

import org.cloudbus.cloudsim.core.CloudSim;

import org.cloudbus.cloudsim.power.PowerHost;

import org.cloudbus.cloudsim.power.models.PowerModelCubic;

import org.cloudbus.cloudsim.provisioners.BwProvisionerSimple;

import org.cloudbus.cloudsim.provisioners.PeProvisionerSimple;

import org.cloudbus.cloudsim.provisioners.RamProvisionerSimple;

import org.cloudbus.cloudsim.power.PowerVmSelectionPolicyMinimumMigrationTime;

import org.cloudbus.cloudsim.power.PowerVmSelectionPolicy;

import org.cloudbus.cloudsim.power.PowerVm;

/**

* @author admin

*/

public class VMAllocation

Details dt=new Details();

Datacenter dc1;

DatacenterCharacteristics characteristics;

public void readVM()

try
{

File fe=new File("vm1.txt");

FileInputStream fis=new FileInputStream(fe);

byte bt[]=new byte[fis.available()];

fis.read(bt);

fis.close();

String g1=new String(bt);

System.out.println("VM List");

System.out.println("=========================");

System.out.println(g1);

String g2[]=g1.split("\n");

for(int i=1;i<g2.length;i++)

dt.Vt.add(g2[i].trim());

dt.vms=new String[dt.Vt.size()][4];

for(int i=0;i<dt.Vt.size();i++)

String a1[]=dt.Vt.get(i).toString().trim().split("\t");

dt.vms[i][0]=a1[0]; // VM Id

dt.vms[i][1]=a1[1]; // VM cpu

dt.vms[i][2]=a1[2]; // VM ram

dt.vms[i][3]=a1[3]; // VM bw
 }

catch(Exception e)

e.printStackTrace();

public void readHost()

try

File fe=new File("host2.txt");

FileInputStream fis=new FileInputStream(fe);

byte bt[]=new byte[fis.available()];

fis.read(bt);

fis.close();

String g1=new String(bt);

System.out.println("Host List");

System.out.println("=========================");

System.out.println(g1);

String g2[]=g1.split("\n");

for(int i=1;i<g2.length;i++)
 {

dt.Ht.add(g2[i].trim());

dt.host=new String[dt.Ht.size()][4];

for(int i=0;i<dt.Ht.size();i++)

String a1[]=dt.Ht.get(i).toString().trim().split("\t");

dt.host[i][0]=a1[0]; // Host Id

dt.host[i][1]=a1[1]; // Host cpu

dt.host[i][2]=a1[2]; // Host ram

dt.host[i][3]=a1[3]; // Host bw

catch(Exception e)

e.printStackTrace();

public void createHost()

try

{
 Log.printLine("Starting CloudSim");

CloudSim cs=new CloudSim();

Calendar calendar = Calendar.getInstance();

cs.init(1, calendar,false);

String name="DC1";

for(int i=0;i<dt.Ht.size();i++)

String a1[]=dt.Ht.get(i).toString().split("\t");

int id=Integer.parseInt(a1[0]);

int cpu=Integer.parseInt(a1[1]);

int ram1=Integer.parseInt(a1[2]);

int bw2=Integer.parseInt(a1[3]);

int storage=100000;

List<Pe> peList1 = new ArrayList<Pe>();

int mips1 = cpu;//1000000;

for(int k=0;k<cpu;k++)

peList1.add(new Pe(0, new PeProvisionerSimple(mips1)));

//peList1.add(new Pe(0, new PeProvisionerSimple(mips1)));

//dt.hostList.add(new Host(id, new RamProvisionerSimple(ram1),new

BwProvisionerSimple(bw2), storage, peList1,new VmSchedulerTimeShared(peList1)));

 dt.hostList.add(new PowerHost(id, new RamProvisionerSimple(ram1),new

BwProvisionerSimple(bw2), storage, peList1,new VmSchedulerTimeShared(peList1),new

PowerModelCubic(1000,500)));

String arch = "x86";

String os = "Linux";

String vmm1 = "Xen";

double time_zone = 10.0;

double cost = 3.0;

double costPerMem = 0.05;

double costPerStorage = 0.2;

double costPerBw = 0.1;

LinkedList<Storage> storageList = new LinkedList<Storage>();

characteristics = new DatacenterCharacteristics(arch, os, vmm1, dt.hostList, time_zone,

cost, costPerMem,costPerStorage, costPerBw);

dc1 = new Datacenter(name, characteristics, new

VmAllocationPolicySimple(dt.hostList), storageList, 0);

System.out.println("Data Center Created with "+dt.Ht.size()+" Host");

catch(Exception e)

{
 e.printStackTrace();

public void createVM()

try

for(int i=0;i<dt.vms.length;i++)

int vmid = Integer.parseInt(dt.vms[i][0]);

int cid=Integer.parseInt(dt.vms[i][0]);

int mips = 250;

long size = 10000; //image size (MB)

int ram = Integer.parseInt(dt.vms[i][2]); //vm memory (MB)

long bw = Long.parseLong(dt.vms[i][3]);

int pesNumber = Integer.parseInt(dt.vms[i][1]); //number of cpus

String vmm = "Xen"; //VMM name

//Vm vm1 = new Vm(vmid,cid, mips, pesNumber, ram, bw, size, vmm, new

CloudletSchedulerTimeShared());

PowerVm vm1 = new PowerVm(vmid,cid, mips, pesNumber, ram, bw, size,1 ,vmm,

new CloudletSchedulerTimeShared(),0.5);
 System.out.println("VM-"+vmid+" is Created...");

dt.vmlist.add(vm1);

catch(Exception e)

e.printStackTrace();

public void optimiseVmAllocation()

try

for(int j=0;j<dt.hostList.size();j++)

PowerHost ph=dt.hostList.get(j);

PSO ps=new PSO();

double uti=ps.fittnessFun(ph);
 System.out.println("Utilization for Host - "+ph.getId()+" = "+uti);

/* for(int i=0;i<dt.vmlist.size();i++)

PowerVm vm=dt.vmlist.get(i);

long vmBW=vm.getBw();

int vmRAM=vm.getRam();

int vmPe=vm.getNumberOfPes();

for(int j=0;j<dt.hostList.size();j++)

PowerHost ph=dt.hostList.get(j);

boolean bool=ph.isSuitableForVm(vm);

if(bool)

int id=ph.getId();

long htBW=ph.getBw();

int htRAM=ph.getRam();

long storage=ph.getStorage();

List<Pe> lt=ph.getPeList();

int htPe=ph.getNumberOfPes();

long bw=htBW-vmBW;
 int ram=htRAM-vmRAM;

int pe=htPe-vmPe;

for(int k=0;k<pe;k++)

lt.add(new Pe(0, new PeProvisionerSimple(pe)));

PowerHost newPH=new PowerHost(id, new RamProvisionerSimple(ram),new

BwProvisionerSimple(bw), storage, lt,new VmSchedulerTimeShared(lt),new

PowerModelCubic(1000,500));

System.out.println(i+" : "+j+" ===== "+ram+" : "+bw+" = "+pe);

//System.out.println(i+" : "+j+" ===== "+htRAM+" : "+newPH.getRam()+" =

"+htBW+" : "+newPH.getBw());

dt.hostList.set(j, newPH);

break;

*/

/* for(int i=0;i<dt.hostList.size();i++)

PowerHost ph=dt.hostList.get(i);
 int id=ph.getId();

long htBW=ph.getBw();

int htRAM=ph.getRam();

long storage=ph.getStorage();

List<Pe> lt=ph.getPeList();

int htPe=ph.getNumberOfPes();

for(int j=0;j<dt.vmlist.size();j++)

PowerVm vm=dt.vmlist.get(j);

boolean bool=ph.isSuitableForVm(vm);

System.out.println(i+" : "+j+" : "+bool+" : "+htBW+" : "+htRAM);//+" :

"+ph.getAvailableMips()+" = "+vm.getMips()+" : "+ph.getNumberOfPes()+" =

"+vm.getNumberOfPes());

if(bool)

long vmBW=vm.getBw();

int vmRAM=vm.getRam();

int vmPe=vm.getNumberOfPes();

long bw=htBW-vmBW;

int ram=htRAM-vmRAM;
 int pe=htPe-vmPe;

for(int k=0;k<pe;k++)

lt.add(new Pe(0, new PeProvisionerSimple(pe)));

PowerHost newPH=new PowerHost(id, new RamProvisionerSimple(ram),new

BwProvisionerSimple(bw), storage, lt,new VmSchedulerTimeShared(lt),new

PowerModelCubic(1000,500));

System.out.println("===== "+newPH.getRam()+" : "+newPH.getBw());

dt.hostList.set(i, newPH);

}*/

catch(Exception e)

e.printStackTrace();

}
10.2 SCREEN SHOTS
VM List

=========================

VMId Cpu Ram BW

1 200 4 50

2 300 7 100

3 150 3 30

4 100 2 50

5 450 6 100

6 300 5 80

7 500 7 150

8 200 3 20

9 100 2 10

10 430 5 50

Host List

=========================

Host_Id CPU RAM Bandwidth

1 1000 20 500

2 1000 10 650

3 500 10 350

4 1000 20 500

5 1000 20 500

6 750 14 650

7 500 10 350

8 1000 20 500

9 600 12 400
10 1000 20 500

Starting CloudSim

Initialising...

Data Center Created with 10 Host

VM-1 is Created...

VM-2 is Created...

VM-3 is Created...

VM-4 is Created...

VM-5 is Created...

VM-6 is Created...

VM-7 is Created...

VM-8 is Created...

VM-9 is Created...

VM-10 is Created...

0.00: VM #1 has been allocated to the host #1

0.00: VM #2 has been allocated to the host #1

0.00: VM #3 has been allocated to the host #1

0.00: VM #4 has been allocated to the host #1

[VmScheduler.vmCreate] Allocation of VM #5 to Host #1 failed by RAM

0.00: Creation of VM #5 on the host #1 failed

VM - 5 is migrated

Utilization for Host - 1 = 246.0

0.00: VM #5 has been allocated to the host #2

[VmScheduler.vmCreate] Allocation of VM #6 to Host #2 failed by RAM

0.00: Creation of VM #6 on the host #2 failed

VM - 6 is migrated

Utilization for Host - 2 = 106.0

0.00: VM #6 has been allocated to the host #3

[VmScheduler.vmCreate] Allocation of VM #7 to Host #3 failed by RAM

0.00: Creation of VM #7 on the host #3 failed

VM - 7 is migrated

Utilization for Host - 3 = 85.0

0.00: VM #7 has been allocated to the host #4

0.00: VM #8 has been allocated to the host #4

0.00: VM #9 has been allocated to the host #4

0.00: VM #10 has been allocated to the host #4

Utilization for Host - 4 = 247.0

Utilization for Host - 5 = 0.0

Utilization for Host - 6 = 0.0

Utilization for Host - 7 = 0.0

Utilization for Host - 8 = 0.0

Utilization for Host - 9 = 0.0

Utilization for Host - 10 = 0.0

76

Range : 1

Range : 1

BUILD SUCCESSFUL (total time: 9 minutes 58 seconds)

 11

11. REFERENCES

[1]. K. Agarwal and T. Kumar, “Combining virtualization and containerization to support

interactive games and simulations on the cloud,” in 2nd International Conference on Intelligent
Computing and Control Systems(ICICCS). IEEE, 2018.

[2]. S. Rajput and A. Arora, “Virtual machine placement in cloud data centers using a hybrid
multi-verse optimization algorithm,” International Journal of Computer Applications, vol. 75, no.
10, pp. 6–12, 2013.

[3]. M. Mohamad and A. Selamat, “An Energy-aware Host Resource Management Framework for
Two-tier Virtualized Cloud Data Centers,” in International Conference on Computer,
Communications, and Control Technology (I4CT). IEEE, 2015, pp. 227–231.

[4]. J. Ramos et al., “An efficient power-aware VM allocation mechanism in cloud data centers: a
micro genetic-based approach,” in Proceedings of the first instructional conference on machine
learning, vol. 242. Piscataway, NJ, 2003, pp. 133–142.

[5]. T. Kumaresan and C. Palanisamy, “An Energy and SLA-Aware Resource Management
Strategy in Cloud Data Centers,” International Journal of Bio-Inspired Computation, vol. 9, no. 3,
pp. 142–156, 2017.

[6]. H. Kaur and S. Ajay, “Renewable Energy-based Multi-Indexed Job Classification and
Container Management Scheme for Sustainability of Cloud Data Centers,” Next Generation
Computing Technologies(NGCT), pp. 516–521, 2016.

[7]. K. Toutanova and C. Cherry, “A placement architecture for a container as a service (CaaS) in
a cloud environment,” in Proceedings of the Joint Conference of the 47th Annual Meeting of the
ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP:
Volume 1- Volume 1. Association for Computational Linguistics, 2009, pp. 486– 494.

[8]. T. N. Sainath, O. Vinyals, A. Senior, and H. Sak, “Energy Consumption Optimization of

Container-Oriented Cloud Computing Center,” in 2015 IEEE International Conference on
Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2015, pp. 4580–4584.
[9]. T. Mikolov and G. Zweig, “An energy, performance efficient resource consolidation scheme
for heterogeneous cloud datacenters,” in 2012 IEEE Spoken Language Technology Workshop
(SLT). IEEE, 2012, pp. 234–239.

[10]. Rizky, W. M., Ristu, S., Afrizal, D. “HeporCloud: An energy and performance efficient
resource orchestrator for hybrid heterogeneous cloud computing environments”. Scientific Journal
of Informatics, Vol. 3(2), p. 41-50, Nov. 2016.