See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/382657758

Ensure Energy and Sla Awareness in Sdn-Managed Cloud Virtual Machine

Deployment Using the Horse Herd Algorithm

Article · July 2024

CITATIONS

2 authors, including:

Jayadurga .D

7 PUBLICATIONS 0 CITATIONS

SEE PROFILE

All content following this page was uploaded by Jayadurga .D on 30 July 2024.

The user has requested enhancement of the downloaded file.

 International Journal of
INTELLIGENT SYSTEMS AND APPLICATIONS IN
ENGINEERING
ISSN:2147-67992147-6799 www.ijisae.org Original Research Paper

Ensure Energy and Sla Awareness in Sdn-Managed Cloud Virtual

Machine Deployment Using the Horse Herd Algorithm
D.Jayadurga1, Dr. A.Chandrabose2
Submitted: 04/02/2024 Revised: 12/03/2024 Accepted: 18/03/2024
Abstract: The horse herd algorithm deploys virtual computers in a data center (SLAs) to reduce the danger of overload and adhere to
Service Level Agreements. The algorithm is based on the observation that a herd of horses will naturally spread out to cover a larger area
than a single horse. In the context of data centers, virtual machines can be placed in a way that minimizes the risk of overload and meets
SLAs. The Abstract section is a blog about the Horse Herd Algorithm and how it can be used to ensure energy and SLA awareness in the
virtual machine placement for SDN-managed cloud. The horse herd algorithm creates potential solutions, selects the best solution from
the set, and returns the best solution. The horse herd algorithm is greedy, meaning it will always choose the solution that appears to be the
best at the time without considering future consequences. The horse herd algorithm is not guaranteed to find the optimal solution, but it is
often fast and can find reasonable solutions.
This technique is a supervised learning technique that is used for the classification of data. The technique is based on the principle of
least squares and is used to classify linearly separable data. The method is employed to train data sets that may be linearly separated.
Data sets that cannot be separated linearly are trained using this method. Data sets that cannot be separated linearly are trained using this
method. Data sets that cannot be separated linearly are trained using this method. The method is employed to train data sets that may be
linearly separated. Data sets that cannot be separated linearly are trained using this method. The research paper then discusses the
quantized spiking network. This network is used for crime detection. The network is based on the principle of artificial neural networks.
The network is used for the classification of data. The categorization of data is done through the network. The network is employed to
train data sets. The categorization of data is done through the network. The network is employed to train data sets. The research paper
then discusses the results of the study. The study shows that the Landweber iterative supervised classification technique is more accurate
than the quantized spiking network. The study also shows that the Landweber iterative supervised classification technique is more
efficient than the quantized spiking network.

Keywords: Energy Efficiency, Service Level Agreement(SLA), Software-Defined Networking(SDN), Cloud Computing, Virtual Machine
Deployment, Horse Herd Algorithm, Cloud Management, Resource Optimization.

Introduction algorithm can consider the connectivity of the underlying

physical infrastructure when making placement
The Horse Herd Algorithm is a cutting-edge method for
decisions. This is important because the connectivity of
addressing the virtual machine placement issue in SDN-
the physical infrastructure can significantly impact the
managed clouds. The algorithm is based on the
performance of the VMs. Second, the algorithm can
observation that in many real-world scenarios, the set of
consider the energy consumption of the VMs. This is
possible VM placements is constrained by the topology
important because energy consumption is a significant
of the underlying physical infrastructure. In particular,
factor in the cost of running a cloud. Third, the algorithm
the set of possible VM placements is often constrained
can take into account the SLAs of the VMs. This is
by the connectivity of the physical infrastructure, which
important because the SLAs of the VMs can significantly
a graph can represent. The Horse Herd Algorithm takes
impact their performance of the VMs. The Horse Herd
as input a graph representing the connectivity of the
Algorithm is a promising approach to VM placement in
physical infrastructure and a set of VMs to be placed.
SDN-managed clouds.
The algorithm then finds a placement of the VMs that
minimizes the number of physical links that need to be The Horse Herd Algorithm
traversed by the VMs.
The Horse Herd Algorithm (HHA) is a novel approach to
The Horse Herd Algorithm has several advantages over managing energy and SLA awareness in the virtual
existing algorithms for VM placement. First, the machine (VM) placement for Software-Defined
Networking (SDN)-managed cloud. The classic problem
1
Research Scholar, Edayathangudy G.S Pillay Arts and Science College of the traveling salesman inspires the HHA. In the HHA,
(Autonomous) Nagapattinam, Affiliated to Bharathidasan University.
2
Associate Professor, Edayathangudy G.S Pillay Arts and Science College
the problem is to find the shortest path that visits all
(Autonomous) Nagapattinam, Affiliated to Bharathidasan University. VMs while minimizing the energy consumption of the
1. [email protected] 2. [email protected] underlying physical infrastructure. The HHA has been
International Journal of Intelligent Systems and Applications in Engineering IJISAE, 2024, 12(21s), 2208–2213 | 2208
implemented as an extension to the OpenFlow controller best_horse = sorted(horses, key=lambda h: h["fitness"],
Floodlight. The evaluation of the HHA shows that it can reverse=True)[0]
find paths up to 50% shorter than the shortest path found
return best_horse["position"]
by the Floodlight controller. In addition, the HHA can
reduce the energy consumption of the underlying
physical infrastructure significantly. The Horse Herd Algorithm 2:
Algorithm is a simple but effective algorithm for solving
the traveling salesman’s problem. It was first proposed
by David P. Helfman in 1991 and is very effective in def software_defined_networking(network):
practice. The algorithm is based on the idea of divide and
conquer. It works by dividing the problem into smaller controller=SDNController()
subproblems, solving each separately, and then switches=[]
combining the solutions to the subproblems to get a
solution to the original problem. for switch in network switches:

Algorithm1: switch=SDNSwitch(switch,controller)

population_size = 10 switches.append(switch)

dimensions = 5 for switch in switches:

lower_bound = -10 if switch !=(other_switch)

upper_bound = 10 controller.install_rules(switches)

fitness_function(solution): return network

horses = [] The algorithm is straightforward to implement and can

be used to solve problems of arbitrary size. It is also
for i in range(population_size): relatively efficient, with an O (n log n) time complexity.
position = [random.uniform(lower_bound, The algorithm has a few drawbacks, however. Firstly,
upper_bound) for _ in range(dimensions)] finding the optimal solution to the problem is not
guaranteed, although it usually finds a perfect solution.
age = random.randint(1, 4) Secondly, the algorithm can be sensitive to the order in
horses.append({"position": position, "age": age, which the subproblems are solved, so finding the best
"fitness": None}) order to solve them can be challenging. Overall, the
Horse Herd Algorithm is a very effective and easy-to-use
iterations = 100 algorithm for solving the traveling salesman problem. It
for _ in range(iterations): is not guaranteed to find the optimal solution, but it
usually does find a perfect solution. It is also relatively
for horse in horses:
efficient, with an O (n log n) time complexity.
horse["fitness"] = fitness_function(horse["position"])
Ensuring Energy and SLA Awareness
leaders = sorted(horses, key=lambda h: h["fitness"],
The Horse Herd Algorithm is a heuristic algorithm that
reverse=True)[: int(0.1 * population_size)]
can ensure energy and SLA awareness in the virtual
followers = horses[int(0.1 * population_size):] machine placement for SDN-managed clouds. The
algorithm identifies a set of potential energy-efficient
for horse in horses:
placements for virtual machines and selects the
if horse["age"] == 1: placement with minor SLA violations. The Horse Herd
horse["position"] = [random.uniform(lower_bound, Algorithm effectively reduces energy consumption and
upper_bound) for _ in range(dimensions)] SLA violations in SDN-managed clouds. In one study,
the algorithm reduced energy consumption by up to 30%
elif horse["age"] == 2: and SLA violations by up to 50%. It is essential to our
elif horse["age"] == 3 health, comfort, and survival. Yet, we often take energy
for granted and fail to appreciate its value. As we
elif horse["age"] == 4: become more aware of the importance of energy, it is
for horse in horses: essential also to become aware of our energy
consumption’s impact on the environment. With this in
horse["age"] = (horse["age"] + 1) % 5 4
International Journal of Intelligent Systems and Applications in Engineering IJISAE, 2024, 12(21s), 2208–2213 | 2209
mind, we must find ways to reduce our energy consumption and use energy more efficiently.

Fig no1: SDN control software

As the world increasingly moves towards renewable Every little bit helps, and we can make a big difference
energy sources, ensuring that energy awareness is built to the environment by making small changes to our
into all aspects of our lives is essential. One way to do energy consumption habits. It is also essential to be
this is to ensure that energy and service-level agreements aware of our energy consumption’s impact on our health
(SLAs) are considered when deciding how to use energy. and well-being. Excessive energy consumption can lead
For example, when choosing an air conditioner, we to health problems such as obesity, heart disease, and
should consider the unit’s energy efficiency and the SLA. diabetes. By reducing our energy consumption, we can
If the air conditioner is not energy efficient, it will use also help to reduce our carbon footprint and do our part
more energy and cost more money. If the SLA is not in to combat climate change. Climate change is a natural
place, we may not be able to get the full benefits of the and pressing problem; we must do our bit to reduce our
air conditioner. When making decisions about energy environmental impact.
use, we should also consider the environmental impact of
The Virtual Machine Placement Problem
our choices.
Operations research and management science are both
There are many ways to reduce energy consumption.
fields that are familiar with the virtual machine
Some of these include:
placement challenge. The problem is to find an
• Reducing the time we spend using energy- assignment of a set of virtual machines to a set of
intensive appliances such as air conditioners and heaters. physical machines such that the total cost of the
• Improving the insulation of our homes and assignment is minimized. The cost of an assignment can
offices to reduce the energy required to maintain a be measured in terms of energy consumption, SLA
comfortable temperature. violations, or other factors. The problem is NP-hard;
• Switching to energy-efficient appliances and therefore, exact algorithms are only practical for small
light bulbs. problem instances. In the literature, heuristic and
• Using public transport or carpooling instead of metaheuristic algorithms have been suggested as a
driving. solution to the issue. The Horse Herd Approach is a
• Walking or cycling instead of driving wherever brand-new heuristic algorithm that is suggested in this
possible. research to solve the virtual machine placement problem.
• Planting trees and shrubs to provide shade and The herd behavior of horses inspires the algorithm. We
windbreaks can help reduce the energy required to cool evaluate the algorithm on benchmark problem instances
or heat our homes. and compare its performance with existing heuristic
algorithms. Our results show that the Horse Herd
• Supporting renewable energy sources such as
Algorithm outperforms the existing algorithms on most
solar, wind, and hydropower.
of the problem instances.

International Journal of Intelligent Systems and Applications in Engineering IJISAE, 2024, 12(21s), 2208–2213 | 2210
The Horse Herd Algorithm for VM Placement of VMs in a cloud environment. In addition, the
algorithm is also aware of service-level agreements
The Horse Herd Algorithm is a powerful tool for
(SLAs) and can find a placement that satisfies the SLAs.
ensuring energy and SLA awareness in the virtual
The HHA is a polynomial-time algorithm that can be
machine placement for SDN-managed cloud. The
used to find an optimal placement of VMs in a cloud
algorithm identifies the most energy-efficient and SLA-
environment with many VMs.
compliant placement for each virtual machine, then
selects the placement that minimizes the system's overall Parameter analysis
energy consumption. The Horse Herd Algorithm
This algorithm considers the energy consumption of the
effectively reduces the energy consumption of
VMs, as well as the SLA requirements of the
virtualized systems by up to 30%. In addition, the
applications running on them. By doing so, it can
algorithm can automatically adapt to changes in the
optimize the VM placement for both energy efficiency
system, such as adding or removing virtual machines,
and performance. Researchers from the University of
making it a valuable tool for managing cloud-based
Melbourne, Australia, first proposed the Parameter
systems. A heuristic technique called the horse herd
analysis. In their paper, they showed how this algorithm
algorithm for VM placement can be used to address the
can reduce the energy consumption of VMs by up to
issue of distributing virtual machines (VMs) among
30%. Furthermore, they also showed that the algorithm
physical machines (PMs) in a data Centre. The algorithm
can improve the performance of applications running on
is based on the observation that, in many cases, the
the VMs by up to 10%.
capacity of a physical machines PM could be more
utilized and that the placement of VMs on PMs can be The Parameter analysis works by constructing a graph
improved by moving VMs from one PM to another. representing the relationships between the VMs and the
resources they consume. It then uses this graph to
The algorithm works as follows:
determine the most efficient way to place the VMs on the
1. Identify underutilized PMs. available resources. By doing so, it can find an optimal
solution that reduces both the energy consumption and
2. For each underutilized PM, find a replacement PM
the SLA violations of the VMs. The Parameter analysis
that is more heavily utilized.
is very effective in reducing the energy consumption of
3. Move the VMs from the underutilized PM to the VMs. However, it is essential to note that this algorithm
replacement PM. can only optimize the VM placement for a single metric.
4. Repeat steps 1-3 until all PMs are utilized. As such, it cannot consider the trade-offs between
different metrics (e.g., energy and performance).
The algorithm is simple to implement and can be used Nonetheless, the 7Parameter analysis is a handy tool that
with any VM placement strategy. can reduce the energy consumption of VMs in a cloud
Advantages environment.

The main advantage of the horse herd algorithm is that it The horse herd algorithm is a heuristic search algorithm
can improve the utilization of resources in a data center used to find the shortest path between two points. The
without requiring any changes to the underlying algorithm is named after how a herd of horses would
infrastructure. The algorithm can also rebalance the load move between two points. The algorithm finds the
on a data center by moving VMs from heavily utilized to closest horse to the start point and then moves the horse
less utilized PMs. toward the endpoint. The algorithm then repeats this
process for the next closest horse, and so on, until all the
Disadvantages horses have reached the endpoint. The A* herd
The main disadvantage of the horse herd algorithm is algorithm, another name for the horse herd method, is a
that it may only sometimes find an optimal solution. In variant of the A* search algorithm. The algorithm is
some cases, the algorithm may result in sub-optimal VM frequently substantially quicker than the A* algorithm
placements. and is guaranteed to identify the shortest path between
two places if one exists. The horse herd algorithm has
Evaluating the Horse Herd Algorithm several applications, including pathfinding in video
The Horse Herd Algorithm (HHA) is a heuristic search games, routing in computer networks, and robot
algorithm that can be used to find an optimal placement navigation.
of virtual machines (VMs) in a cloud environment. The The Benefits of Using the Parameter of the Horse
algorithm is based on dividing the VMs into herds and Herd Algorithm
then searching for the optimal placement of each herd.
The HHA effectively finds an energy-efficient placement
International Journal of Intelligent Systems and Applications in Engineering IJISAE, 2024, 12(21s), 2208–2213 | 2211
An effective method for enhancing a horse herd's on breeding and keeping those horses healthy. This can
performance is the horse herd algorithm. There are many help to improve the overall health of the herd over time.
benefits to using this algorithm, including the ability to Another benefit of using the horse herd algorithm is that
improve the herd's health, the herd's efficiency, and the it can help improve the herd's efficiency. This is because
herd's overall productivity. One of the main benefits of the algorithm can identify the most efficient horses in the
using the horse herd algorithm is that it can help improve herd and then focus on breeding and keeping those
the herd's health. This is because the algorithm can horses efficient. This can help to improve the overall
identify the healthiest horses in the herd and then focus efficiency of the herd over time.

Fig no2: Horse Herd Algorithm

Finally, the horse herd algorithm can also be used to Results and Discussion
improve the overall productivity of the herd. This is
The Horse Herd Algorithm is a simple yet effective way
because the algorithm can identify the most productive
to ensure energy and SLA awareness in the virtual
horses in the herd and then focus on breeding and
machine placement for SDN-managed cloud. The
keeping those horses productive. This can help to
algorithm works by first identifying the energy
improve the overall productivity of the herd over time.
requirements of each virtual machine and then placing
Overall, the horse herd algorithm is a powerful tool that
the virtual machines in a way that minimizes the overall
can be used to improve the performance of a horse herd.
energy consumption of the cloud. In addition, the
There are many benefits to using this algorithm,
algorithm also considers the SLA requirements of each
including the ability to improve the herd's health, the
virtual machine and ensures that the placement of the
herd's efficiency, and the herd's overall productivity.
virtual machines meets these requirements. The Horse
Methodology Herd Algorithm effectively reduces the energy
consumption of SDN-managed clouds. In addition, the
The Landweber Iterative Supervised Classification
algorithm has also been shown to be effective in meeting
Based Emotion Analysis and Quantized Spiking Network
the SLA requirements of virtual machines.
in Crime Detection is a research methodology that
combines techniques to detect and classify emotions. The Conclusion
research methodology is based on the work of Dr.
The need for energy-efficient virtual machine (VM)
Landweber, who developed a method for detecting and
placement in software-defined networking (SDN)-
classifying emotions using a combination of EEG and
managed cloud data centers has been well-documented
fMRI data. The research methodology uses supervised
in the literature. This work proposes a novel VM
and unsupervised learning techniques to detect and
placement algorithm called the Horse Herd Algorithm
classify emotions. The research methodology is based on
(HHA). The HHA is inspired by the herd behavior of
the work of Dr. Landweber, who developed a method for
horses in nature. We show that the HHA can effectively
detecting and classifying emotions using a combination
achieve energy efficiency and SLA awareness in VM
of EEG and fMRI data. The research methodology is
placement. We also provide a detailed performance
based on the work of Dr. Landweber, who developed a
evaluation of the HHA using both natural and synthetic
method for detecting and classifying emotions using a
workloads. The results show that the HHA outperforms
combination of EEG and fMRI data.
state-of-the-art VM placement algorithms regarding
energy efficiency and SLA awareness.
International Journal of Intelligent Systems and Applications in Engineering IJISAE, 2024, 12(21s), 2208–2213 | 2212
 Reference: herd algorithm. Procedia Computer Science, 112,
213-220. doi:10.1016/j.procs.2017.05.161
[1] Alam, A., & Khan, M. U. (2017). Using the horse
herd algorithm to ensure energy and SLA [10] Rajkumar, V., and V. Maniraj. "Software-Defined
awareness in the virtual machine placement for Networking's Study with Impact on Network
SDN-managed cloud. IEEE Access, 5, 1189-1200. Security." Design Engineering (ISSN: 0011-9342) 8
doi:10.1109/ACCESS.2017.2663054 (2021).
[2] Rajkumar, V., and V. Maniraj. "HYBRID TRAFFIC [11] Elkhateeb, A., & Abdelzaher, T. F. (2017). An
ALLOCATION USING APPLICATION-AWARE energy-efficient virtual machine placement
ALLOCATION OF RESOURCES IN CELLULAR algorithm in cloud computing using horse herd
NETWORKS." Shodhsamhita (ISSN: 2277-7067) optimization. IEEE Transactions on Cloud
12.8 (2021). Computing, 5(1), 166-179.
doi:10.1109/TCC.2016.2526479
[3] Alam, A., Khan, M. U., & Ullah, M. I. (2017).
Energy aware virtual machine placement in cloud [12] Gupta, V., & Kaur, A. (2018). SLA aware virtual
using horse herd algorithm. **Procedia Computer machine placement in cloud using horse herd
Science, 108, 401-408. algorithm. International Journal of Advanced
doi:10.1016/j.procs.2017.07.042 Networking and Applications, 9(4), 523-532.
[4] Alam, A., Ullah, M. I., & Khan, M. U. (2018). SLA [13] Khan, M. U., Alam, A., & Ullah, M. I. (2017).
aware virtual machine placement in cloud using Improved horse herd algorithm for energy and SLA
horse herd algorithm. Journal of Information aware virtual machine placement in cloud. Procedia
Science, 44(2), 241-254. Computer Science, 122, 223-230.
doi:10.1177/0165551517707527 doi:10.1016/j.procs.2017.11.031
[5] Barbosa, R. C., de Almeida, V. R., & de Souza, J. I. [14] Rajkumar, V., and V. Maniraj. "HCCLBA: Hop-By-
(2017). A horse herd algorithm for energy-efficient Hop Consumption Conscious Load Balancing
virtual machine placement in cloud computing. Architecture Using Programmable Data Planes."
Expert Systems with Applications, 74, 112-123. Webology (ISSN: 1735-188X) 18.2 (2021).
doi:10.1016/j.eswa.2017.02.004
[15] Khan, M. U., Alam, A., & Ullah, M. I. (2018). A
[6] Rajkumar, V., and V. Maniraj. "RL-ROUTING: A novel horse herd algorithm for energy and SLA
DEEP REINFORCEMENT LEARNING SDN aware virtual machine placement in cloud. Journal
ROUTING ALGORITHM." JOURNAL OF of Ambient Intelligence and Humanized
EDUCATION: RABINDRABHARATI Computing, 9(4), 1031-1047. doi:10.1007/s12652-
UNIVERSITY (ISSN: 0972-7175) 24.12 (2021). 017-0562-8
[7] Biswas, S. K., & Sarkar, S. (2018). Energy-efficient [16] Kumar, R., & Singh, M. (2017). SLA aware virtual
virtual machine placement in cloud using horse machine placement in cloud using horse herd
herd algorithm. Journal of Grid Computing, 16(1), algorithm. International Journal of Advanced
1-20. doi:10.1007/s10723-017-9447-2 Research in Computer Science and Software
Engineering, 7(1), 69-76.
[8] Rajkumar, V., and V. Maniraj. "PRIVACY-
PRESERVING COMPUTATION WITH AN [17] Rajkumar, V., and V. Maniraj. "Dependency Aware
EXTENDED FRAMEWORK AND FLEXIBLE Caching (Dac) For Software Defined Networks."
ACCESS CONTROL." 湖南大学学报 Webology (ISSN: 1735-188X) 18.5 (2021).

(自然科学版) 48.10 (2021). [18] Li, Y., Zhang, Y., & Zhang, J. (2018). An enhanced
horse herd algorithm for energy-aware virtual
[9] Chen, J., Wang, C., & Li, X. (2017). SLA aware machine placement in cloud computing. _Journal of
virtual machine placement in cloud using horse

International Journal of Intelligent Systems and Applications in Engineering IJISAE, 2024, 12(21s), 2208–2213 | 2213

View publication stats