Virtual machine placement in cloud using artificial bee colony and imperialist competitive algorithm

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 13, No. 4, August 2023, pp. 4743~4751
ISSN: 2088-8708, DOI: 10.11591/ijece.v13i4.pp4743-4751  4743
Journal homepage: http://ijece.iaescore.com
Virtual machine placement in cloud using artificial bee colony
and imperialist competitive algorithm
Seyyed-Mohammad Javadi-Moghaddam1
, Zahra Dehghani2
1
Department of Computer, Faculty of Engineering, Bozorgmehr University of Qaenat, Qaen, Iran
2
Department of Computer Engineering, Azad University of Birjand, Birjand, Iran
Article Info ABSTRACT
Article history:
Received Sep 14, 2022
Revised Dec 8, 2022
Accepted Dec 11, 2022
Increasing resource efficiency and reducing energy consumption are
significant challenges in cloud environments. Placing virtual machines is
essential in improving cloud systems’ performance. This paper presents a
hybrid method using the artificial bee colony and imperialist competitive
algorithm to reduce provider costs and decrease client expenditure.
Implementation of the proposed plan in the CloudSim simulation environment
indicates the proposed method performs better than the Monarch butterfly
optimization and salp swarm algorithms regarding energy consumption and
resource usage. Moreover, average central processing unit (CPU) and
random-access memory (RAM) usage and the number of host shutdowns
show better results for the proposed model.
Keywords:
Artificial bee colony algorithm
Cloud computing
Imperialist competitive
algorithm
Virtual machine placement
Virtualization
This is an open access article under the CC BY-SA license.
Corresponding Author:
Seyyed-Mohammad Javadi-Moghaddam
Department of Computer, Faculty of Engineering, Bozorgmehr University of Qaenat
Abolmafakher Street, Qaen, Iran
Email: smjavadim@buqaen.ac.ir
1. INTRODUCTION
A particular distributed computing model that introduces applicable models for providing remotely
measurable resources and commercial computing network commands, parallel computing, and distributed
computing is known as cloud computing [1], [2]. The demand for users for cloud resources is increasing while
cloud computing technology is developing [3]. The increase in the resources provided in the cloud causes an
increase in essential parameters such as the amount of energy consumption, which can cause new challenges
such as limited energy resources and global warming.
Technology virtualization is essential to improve the efficient use of clouds and resources, which has
many advantages, including server integration, migration, isolation accessibility, flexible expansion, and low
management overhead. Virtualization offers many energy management solutions in cloud computing [4], [5].
The virtualization process can integrate several virtual machines (VM) at the data center level by placing some
virtual servers on a physical machine (PM) and then shutting down idle PMs to optimize energy consumption
[6], [7]. The virtual machines placement (VMP) places the VMs on the physical machines so that the benefit
of the cloud providers is maximized [8]. Each PM has limited resources to host several virtual machines, and
each has its resource requests. The purpose of the VMP problem is to place the maximum possible number of
VMs on the PMs while not exceeding the physical range of each physical machine and resulting in maximum
profit for the provider [9].
Various methods have been proposed for locating machines in cloud computing [10], [11]. However,
they are still facing the challenge of increasing the energy consumption related to data centers, raising costs,
and the high execution time of tasks [12]. Huang et al. [13] proposed a suitable destination for selecting virtual

 ISSN: 2088-8708
Int J Elec & Comp Eng, Vol. 13, No. 4, August 2023: 4743-4751
4744
machines by classifying workflows and integrating physical machines. Alharbi et al. [14] presented a linear
formula for the problem of dynamic integration of VMs and limited the mapping of some VMs to private
servers with particular features. They also set a threshold for the number of migrations that do not exceed that
limit. Furthermore, they considered cost parameters in their formulation so that the VM is positioned to reduce
the overall cost of the system. Han et al. [15] described an efficient energy resource management system for
virtual cloud centers, which reduces operating costs and satisfies the quality of service as much as possible.
Sharma et al. [16] proposed an approach for the VMP problem to improve cloud service quality. This work
tries to use fewer hosts to reduce the overhead. Braiki and Youssef [17] have introduced a self-adaptive solution
for the VMP algorithm using particle swarm optimization. Its self-adaptive feature allows the algorithm to
decide when each virtual machine is placed on which physical host.
Some researchers have focused on only energy saving [18], [19]. Kaaouache and Bouamama [20]
proposed an approach to solving the problem using genetic algorithms. This approach attempts to optimize the
energy consumption in servers and communications with a suitable objective function. Reddy et al. [21]
suggested a stabilization technique to reduce energy consumption in a cloud system. The authors have
empirically described server power consumption as a central processing unit (CPU) performance and disk
usage model. Dhaya et al. [22] focused on energy management methods of resources that a cloud provider can
process in a virtual data center. A primary tool is live VM migration. The ability of VMs to migrate between
low-overhead physical hosts provides the necessary flexibility for the resource provider to dynamically
reallocate virtual machines according to current resource requirements and allocation policies. This method
shuts idle physical nodes down to minimize energy consumption and optimizes placement.
Several articles have considered the service level agreement (SLA). Alharbi et al. [23], try to minimize
SLA violations number using the ant colony algorithm. In this multipurpose algorithm, the behavior of each
ant is a solution for assigning the virtual machine to the server. Solutions created by specific functions are then
evaluated. If this solution is on the list of the best solutions, its pheromone level will increase. Otherwise, it
will decrease. Finally, the law of updating the pheromone is applied to achieve a suitable solution for virtual
machine placement. Dynamic placement algorithms can improve the integration of servers by reducing the
violation SLA rate. Ranjbari and Torkestani [24] presented a dynamic placement method of virtual machines
to manage breaches of service level parameters. This solution aims to decrease the needed capacity and the
SLA violation rate. The management algorithm presented in this paper reduces the physical ability needed to
support a specific service level violation rate. This algorithm aims to achieve the minimum cost of the data
center implementation based on capacity overflow, low resource consumption, and load overflow, which leads
to poor performance and breach of service level agreement.
Most previous works have tried to decrease the energy consumption of cloud sources by considering
the energy crisis and global warming issues. Solutions have been presented by reducing the number of hosts
and using virtualization to increase the productivity of physical resources. The main challenge in placing a
virtual machine is choosing the suitable virtual machine and physical machine. Meta-heuristic methods have
shown acceptable performance in the search process and are a solution to this problem. This paper attempts to
overcome this challenge using the artificial bee colony algorithm (ABCA) and the imperial competitive
algorithm (ICA). This paper performs the VMP in two stages: select a virtual machine to transfer to physical
servers and select the most appropriate physical server. The proposed model uses the ABCA to choose the best
VM and the imperial competitive algorithm to select the best physical server.
The continuation of the article includes the following sections. Section 2 explains the proposed
method in detail. The evaluation results and discussion have been expressed in section 3. The conclusion is
presented in section 4.
2. METHOD
2.1. Artificial bee colony algorithm (ABCA)
Akay et al. [25] introduced the ABCA algorithm to optimize mathematical functions. Each solution
expresses an area of the potential food area. Furthermore, the quality is the quality of the food source. Artificial
bees try to find food resources by searching. ABCA uses three types of agents: employed bees (EB), onlooker
bees (OB), and scouts. EBs are related to the current answers of the algorithm. At each step of the algorithm,
an EB tries to provide a solution, improve it by a local search step, and find the OB bees for the current location.
OBs select higher quality locations from enhanced locations. If OB could find a better location, EB would
update its location; otherwise, it remains in its current location. Furthermore, an EB bee will leave its place if
it cannot improve its location in a certain number of steps. If an EB leaves its place, it becomes a scout and
randomly selects a new place in the search space.

Int J Elec & Comp Eng ISSN: 2088-8708 
Virtual machine placement in cloud using artificial bee colony … (Seyyed-Mohammad Javadi-Moghaddam)
4745
2.2. Imperialist competitive algorithm (ICA)
The imperial competitive algorithm (ICA) [26] is a proposed optimization method inspired by
imperialist competition's mathematical modeling. This algorithm is an optimization strategy mimicking human
behavior in a social-political evolution. It starts with a random number of initial populations known in a
country. Imperialists united these colonies with the policy of attraction regarding their power. The empire's
power is calculated according to the imperialist state and its colonies. Mathematically, the dependence is
defined according to the total imperial state power and the average of its colonies' power. Imperialism started
with the forming of primitive empires. Any unsuccessful empire in the colonial competition that decreased its
power will be removed from the colonial competition. Thus, the empire's stability is related to its power to
absorb and subjugate the colonies. Therefore, the weaker empires gradually are removed during the imperialist
rivalries because of the enormous empire's power. Empires must increase their power using developing their
colonies. Thus, the power of the colonies will be closed to the empires over time, and convergence will appear.
The algorithm’s end is when an empire’s location is very close to the imperialist country.
2.2.1. Absorption policy
In the absorption phase, the imperialist countries build infrastructures (transportation infrastructure,
university establishment, and to increase their development. This phase of the imperialist optimization
algorithm models the movement of the colonies. Figure 1 depicts this movement. Colony country has moved
X units in the direction of the line connecting the colony to the Imperialist and has changed the position in the
colony. The distance between the imperialist and the colony is denoted by d. A suitable distribution, like
uniform distribution, randomly generates the value of X.
Figure 1. The move of the colonies toward the imperialist [27]
2.2.2. Absorption policy exchanging the colony and imperialist position
Some colonies can take a more suitable situation than the imperialists as the colonies move toward
the colonial country. This situation can be caused by changing the place of the imperialist country and the
colonized country. Then, the algorithm continues on the new position of the colonizing country, and the
colonized countries move to a new place Figure 2.
Figure 2. Exchanging the colony and imperialist position [28]
2.3. Proposed algorithm
The proposed method consists of two parts: selecting a virtual machine to transfer to physical servers
and selecting the most appropriate physical server. It uses the bee colony method to choose the best virtual
machine. Furthermore, the imperialist competitive algorithm tries to select the best physical server.

 ISSN: 2088-8708
4746
2.3.1. Selecting a VM with the ABCA
The ABCA operates using the population criterion, like the exploratory behavior of bees. The ABCA
includes three groups EB, OB, and scouts. Food resource’s location and an available solution to choose the
best virtual machine for placement in machines are considered. The usefulness of a food source indicates the
degree to which it fits the solution to the problem, which is the choice of the best VM in this work. The amount
of EB equals the amount of OB and the number of food sources. Scouts replace any food source with a new
one when it is no further improved in a given cycle.
At first, the initial population is the number of solutions (Ns) that are randomly generated. 𝑁s is equal
to EB and is the food sources number. Furthermore, Ns is the population size that each solution 𝑥𝑖 (𝑖 = 1. . . 𝑁𝑠)
is a multidimensional vector. The bees then conduct a rotating search according to specific rules. In this study,
each EB considers several VM as new candidates as a food source to update the possible solutions. Selection
is made using the previous neighborhood of VMs. As shown in (1) generates candidate solution SVi (solution
i) by previous solutions as (1),
𝑆𝑉𝑖𝑗 = 𝑥𝑖𝑗 + 𝛷𝑖𝑗 (𝑥𝑖𝑗 + 𝑥𝑘𝑗) (1)
where 𝑘 ∈ {1. . . 𝑁𝑠} and 𝑗 ∈ {1. . . 𝑛} are indices randomly generated by a random distribution in the interval
[-1,1]. The method compares candidate solutions with the old ones, and if the new sources are of better quality
or equal to the old food sources, the old food sources will be replaced with new food sources. Otherwise, the
old food sources will remain without any change. After returning ΕΒ to the hive, it shares information about
the food sources information with the ΟΒ. Then, each OB chooses a food source regarding its fitness value. As
shown in (2) calculates the probability 𝑝𝑖 of each selection:
𝑝𝑖 =
𝑓𝑖
∑ 𝑓𝑗
𝑁𝑠
𝑗=1
(2)
where 𝑓𝑖 expresses the fitness value of solution i which depends on the value of the fitness functions, the degree
of fitness can increase the possibility of choosing a food source for each OB. After selecting the food source,
each OB can find a new food source candidate from their neighbors. A greedy solution selects a food source
regarding its fitness. If the position of a food source (optimal virtual machine) is not changed by a finite cycle,
it is considered a solution. Scouts play an essential role in replacing food sources. Scouts replace any old food
source with a new food source. A cycle limitation is a predefined number that the user defines. If source xi
wants to be replaced by scouts (3) is used.
𝑥𝑖𝑗 = 𝑥𝑗𝑚𝑖𝑛 + 𝑟𝑎𝑛𝑑[0.1](𝑥𝑗𝑚𝑎𝑥 − 𝑥𝑗𝑚𝑖𝑛) (3)
where 𝑥𝑗𝑚𝑎𝑥 and 𝑥𝑗𝑚𝑖𝑛 are the bounds of 𝑥𝑖𝑗, a random number between [0.1] is generated by a uniform
distribution. This process is repeated until the function reaches its threshold. Therefore, the bee colony
algorithm determines the best VM placed in the PM. One of the essential operators used in the bee colony
algorithm is the fitting operator, which with the help of other operators, produces the most optimal solutions
to select the desired virtual machine. The process of transmitting messages in the cloud computing network is
that the nodes in each physical machine send their message to the host, and the host also sends its messages to
the data center. As shown in (4) transfers a message from a VM to a physical one after placement.
𝐸𝑛 = ∑ 𝐸𝑛𝑉𝑀𝑖 + ((𝑛 − 1) ∗ 𝐸𝑛𝑃𝐻) + 𝑉𝑀𝑀 + 𝐸𝑃𝑅
𝑛
𝑖=1 (4)
where 𝐸𝑛 refers to the energy used to transmit a message from a virtual machine to a physical machine, n is
the number of virtual machines of the physical machine, and 𝐸𝑛𝑉𝑀𝑖 is the energy consumed to send a message
from the 𝑉𝑀 to the 𝑃𝑀. The 𝐸𝑛𝑃𝐻 parameter is the amount of energy the physical machine requires to receive
the message. 𝐸𝑃𝑅 indicates a physical machine's energy to process and aggregate data. Parameter 𝑉𝑀𝑀
expresses the energy needed to process requests by virtual machines.
Two critical factors play a role in calculating the fitness function: the amount of data overhead and
the energy consumed to send information from the virtual machine to the data center. These two factors are
very influential in choosing the best virtual machine. Therefore, the principal purpose of the ACBA is to
minimize these factors to choose the best VM for placement in the physical machine. Furthermore, we can
reduce the VMs number in the fitting function, which, like reducing the amount of data overhead, can affect
the choice of virtual machines because the host consumes more energy than others. Consequently, the fitness
function consists of a mapping from a virtual machine and a physical machine to bits of zero, and one is
calculated as (5).

4747
𝑓𝑖 =
1
𝐸𝑛
+ (𝑂𝑣𝑒𝑟𝐹 − 𝑆𝑢𝑚𝐸𝑛) + (𝑁 − 𝑃𝐻𝑀) (5)
where 𝑁 is the total number of virtual machines in the cloud computing network, 𝑃𝐻𝑀 determines the PMs
number, and 𝑂𝑣𝑒𝑟𝐹 determines the total data overhead of all VMs. Finally, SumEn specifies the total energy
required to process the virtual machines.
Less energy consumption and less data overload will cause more amount of fitness. The proposed bee
colony algorithm tries to find a suitable solution by increasing the fitting value. Therefore, all the mentioned
operators are used to improve the positioning and calculate the optimal response. After choosing the best VMs
for placement in the physical machine, the method chooses the best.
2.3.2. Selecting a PM with the ICA
The parameters of the ICA include the following: i) Initial population: A number of these countries
are considered primitive populations (Nc). The country describes physical machines; ii) Selection: 𝑁𝐼 is to be
chosen as imperialist from the best members of this population (countries with a minor cost function);
iii) Remaining: 𝑁𝑐𝑜𝑙 expresses the remaining countries which belong to the system. For example, the physical
machines belong to the data center; and iv) Colonies division: Each imperialist is assigned several colonies,
the number commensurate with its power. The normalization cost is considered according to the cost of all
imperialists, as (6).
𝐶𝑛 = 𝑚𝑎𝑥𝑖{𝑐𝑖} − 𝑐𝑛 (6)
where 𝑐𝑛 is the cost of the ith
imperialist. 𝑚𝑎𝑥𝑖{𝑐𝑖} is the maximum cost of imperialists, and 𝐶𝑛 is the
normalization cost. Stronger imperialists will cost more, and weaker imperialists will have fewer normalization
costs. As shown in (7) calculates the relative normalized power of each imperialist (𝑃𝑛).
𝑃𝑛 = |
𝐶𝑛
∑ 𝑐𝑖
𝑁𝐼
𝑖=1
| (7)
The proportion of colonies ruled by an imperialist is the normalized power of an imperialist. Thus, the
initial number of colonies of an imperialist would be equal to (8).
𝑁𝑖𝑛𝑖𝑡. 𝐶𝑙 = 𝑟𝑜𝑢𝑛𝑑(𝑃𝑛. 𝑁𝑐𝑙) (8)
where 𝑁𝑖𝑛𝑖𝑡. 𝐶𝑙 is the initial number of colonies of a system. 𝑁𝑐𝑙 is the number of colonized countries of early
countries (physical machines). round expresses the closest integer to a decimal number. Each system randomly
selects the original colonial countries and assigns them to the nth
imperialist. The colonial competition
algorithm begins after calculating the initial state of all systems. The evolutionary process continues until a
stop condition is met.
2.3.3. Goals of the proposed algorithm
This research pursues two general goals for locating VM in data centers of the cloud system. First,
minimizing overall traffic by placing virtual machines in cloud data centers can reset the overall traffic layout
between VMs and allow VMs to be on the same physical machine or switch with the same amount of traffic.
Second, reducing the maximum link utilization (MLU) enables network traffic to be allocated fairly and avoids
congestion. Moreover, the above criteria consider the quality of service for applications sensitive to latency.
a. Optimizing overall traffic
Two matrices are considered to optimize the overall traffic: the traffic matrix 𝐴 = (𝑎𝑖𝑗)𝑛𝑣×𝑛𝑣 and the
transmission cost matrix 𝐵 = (𝑏ℎ𝑝)𝑛𝑝×𝑛𝑝. 𝑎𝑖𝑗 provides traffic between 𝑉𝑀𝑖 and 𝑉𝑀𝑗, 𝑏ℎ𝑝 provides 𝑃𝑀ℎ and
𝑃𝑀𝑝 transmission costs, nv is the number of VMs and np is the number of PMs. The transmission cost expresses
the switches number that traffic between PM. The greater the number of switches results in the higher the
transmission cost. The goal is to find a mapping function π(i) to encounter VMi located on the PM so that a
VM only can be on one PM, but one PM can be a host for multiple VMs. As shown in (9) calculates the
objective function.
𝑀𝑖𝑛𝑖𝑚𝑖𝑧𝑒 𝑇𝑐𝑜𝑠𝑡 = ∑ 𝑎𝑖𝑗𝑏(𝑖)(𝑗) + ∑ 𝑒𝑖𝑔(𝑖)
𝑛𝑣
𝑖=1
𝑛𝑣
𝑖,𝑗=1 (9)
where 𝑒𝑖 is the traffic between 𝑉𝑀𝑖 and the external communication of the data center, gπ(i) is the transmission
cost from the PM of 𝑉𝑀𝑖 to the external switch.

 ISSN: 2088-8708
4748
b. MLU optimization
The goal of optimization is to minimize the MLU value, which is calculated as (10).
𝑀𝑖𝑛(𝑀𝐿𝑈) = 𝑀𝑎𝑥(
∑ 𝑇𝑟𝑖𝑗
𝑠𝑡
(𝑠,𝑡)
𝐶𝑃𝑖𝑗
) (10)
where 𝑇𝑟𝑖𝑗
𝑠𝑡
is traffic flow from node s to node t traversing link from node i to node j. 𝐶𝑃𝑖𝑗 is the network
capacity of a path from s to t.
3. RESULTS AND DISCUSSION
3.1. Simulation Environment
This study has used CloudSim Simulator [29], [30] to evaluate the proposed method. The operating
system was a 32-bit Windows 7. Features of the simulation hardware of the computer include 4 GB RAM, usable
memory is 3.06 GB, Intel processor (Q 720 @ 1.60GHz 1.60 GHz – (CoreTM) i7 CPU).
3.2. Parameters
This section expresses the characteristics of the simulation environment. Table 1 shows the essential
parameters and fundamental arguments for simulating the proposed method. Furthermore, Table 2 describes
the parameters of the optimization algorithm.
Table 1. The proposed system simulation input parameters and arguments
Parameters and arguments Value
Number of virtual machines 256
Number of hosts 64
Minimum number of processors for each virtual machine 1
Maximum number of processors for each virtual machine 8
Minimum amount of RAM for each virtual machine 1024
Maximum amount of RAM for each virtual machine 8192
Minimum amount of disk space for each virtual machine (GB) 1 GB
Maximum amount of disk space for each virtual machine (GB) 8 GB
Minimum number of processors for each host 1
Maximum number of processors for each host 32
Minimum amount of RAM for each host 1024
Maximum amount of RAM for each host 32768
Minimum amount of disk space per host (GB) 1 GB
Maximum amount of disk space per host (GB) 300 GB
Table 2. Optimization algorithm parameters
Initial population 265
Number of particles 64
Number of iterations of the algorithm 100
3.3. Results
For evaluation, the proposed method, Monarch butterfly optimization, and salp swarm algorithms
have been compared based on energy, CPU, and RAM consumption criteria. Figure 3 shows the three
algorithms’ energy consumption with working loads from 100 to 400 VM. The proposed method with low
energy consumption reduces the cost for cloud service providers. High energy consumption related to data
centers results in increasing costs for cloud computing service providers. Furthermore, it causes increasing CO2
emissions. Figure 3 shows that the proposed method consumed less energy than the compared methods. This
energy consumption reduction states that using two meta-heuristic algorithms in choosing the appropriate
virtual and physical machine has reduced the number of hosts and ultimately reduced energy consumption.
Figure 4 compares the average CPU and RAM usage of the methods. Figure 4 shows that CPU usage
is approximately 95% for the proposed solution, while 90% in the Monarch butterfly optimization and 84% in
the salp swarm algorithm. RAM usage is about 93% for the proposed algorithm, approximately 85% for
monarch butterfly optimization, and 80% for the salp swarm algorithm. Furthermore, it results in the proposed
algorithm increasing resource usage significantly.
Figure 5 depicts the host shutdowns number of the evaluated three methods. Furthermore, it expresses
that the total shutdown of the host in the proposed solution compared to other solutions is much higher.

4749
Consequently, the proposed solution saves a significant percentage of all physical machines. Therefore, the
proposed algorithm solution effectively improves the rate of resource usage by increasing the host capacity and
shutting down more physical machines.
Figure 3. Energy consumption with workloads from 100 VM to 400 VM by the proposed algorithm, Monarch
butterfly optimization, and salp swarm algorithms
Figure 4. Average CPU and RAM usage for the proposed and compared methods
Figure 5. Comparison of the number of hosts shut down in the proposed method and methods compared
Int J Elec & Comp Eng, Vol. 99, No. 1, Month 2099: 1-1x
Swarm algorithms have been compared based on energy, CPU, and RAM consumption
criteria. Figure 3 shows the three algorithms’ energy consumption with working loads from
100 to 400 VM. The proposed method with low energy consumption reduces the cost for
cloud service providers. High energy consumption related to data centers results in
increasing costs for cloud computing service providers. Furthermore, it causes increasing
CO2 emissions. Figure 3 shows that the proposed method consumed less energy than the
compared methods. This energy consumption reduction states that using two meta-
heuristic algorithms in choosing the appropriate virtual and physical machine has reduced
the number of hosts and ultimately reduced energy consumption.
Figure 3. Energy consumption with workloads from 100 VM to 400VM by the proposed
algorithm, Monarch Butterfly Optimization, and Salp Swarm algorithms
Figure 4 compares the average CPU and RAM usage of the methods. Figure 4
shows that CPU usage is approximately 95% for the proposed solution, while 90% in the
Monarch Butterfly Optimization and 84% in the Salp Swarm algorithm. RAM usage is
about 93% for the proposed algorithm, approximately 85% for Monarch Butterfly
Optimization, and 80% for the Salp Swarm algorithm. Furthermore, it results in the
proposed algorithm increasing resource usage significantly.
Figure 5 depicts the host shutdowns number of the evaluated three methods.
Furthermore, it expresses that the total shutdown of the host in the proposed solution
compared to other solutions is much higher. Consequently, the proposed solution saves a
significant percentage of all physical machines. Therefore, the proposed algorithm solution
effectively improves the rate of resource usage by increasing the host capacity and shutting
down more physical machines.
Finally, by reducing the number of hosts, the proposed method can reduce energy
consumption and thus reduce the cost of cloud service providers. This improvement allows
cloud services to be offered more reasonably priced for users.
Workload
Methods
Methods

 ISSN: 2088-8708
4750
Finally, by reducing the number of hosts, the proposed method can reduce energy consumption.
Consequently, it reduces the cost of cloud service providers. This improvement allows cloud services to be
offered more reasonably priced for users.
4. CONCLUSION
An essential goal of cloud environments is to decrease resource provider and client costs. The
proposed model uses the bee colony algorithm to select the best virtual machine and the imperial competition
algorithm to choose the best physical server. The evaluation results depicted that the proposed algorithm has
an average CPU usage of approximately 95% and average RAM usage of about 93%, which is more than both
Monarch butterfly optimization and salp swarm algorithms. Furthermore, the total shutdown of the host in the
proposed solution compared to other solutions is much higher. Therefore, the proposed solution saves a
significant percentage of usage of all physical machines.
REFERENCES
[1] D. A. Shafiq, N. Z. Jhanjhi, and A. Abdullah, “Load balancing techniques in cloud computing environment: A review,” Journal of
King Saud University - Computer and Information Sciences, vol. 34, no. 7, pp. 3910–3933, Jul. 2022, doi:
10.1016/j.jksuci.2021.02.007.
[2] M. Kumar and S. C. Sharma, “PSO-based novel resource scheduling technique to improve QoS parameters in cloud computing,”
Neural Computing and Applications, vol. 32, no. 16, pp. 12103–12126, Aug. 2020, doi: 10.1007/s00521-019-04266-x.
[3] M. Kumar, A. Kishor, J. Abawajy, P. Agarwal, A. Singh, and A. Y. Zomaya, “ARPS: An autonomic resource provisioning and
scheduling framework for Cloud platforms,” IEEE Transactions on Sustainable Computing, vol. 7, no. 2, pp. 386–399, Apr. 2022,
doi: 10.1109/TSUSC.2021.3110245.
[4] A. Rashid and A. Chaturvedi, “Virtualization and its role in cloud computing environment,” International Journal of Computer
Sciences and Engineering, vol. 7, no. 4, pp. 1131–1136, Apr. 2019, doi: 10.26438/ijcse/v7i4.11311136.
[5] R. Rastogi and N. Aggarwal, “A review on virtualization and cloud security,” in 2022 2nd International Conference on Innovative
Practices in Technology and Management (ICIPTM), Feb. 2022, pp. 162–166, doi: 10.1109/ICIPTM54933.2022.9754172.
[6] S. Atiewi, A. Abuhussein, and M. A. Saleh, “Impact of virtualization on cloud computing energy consumption,” in Proceedings of
the 2nd International Symposium on Computer Science and Intelligent Control, Sep. 2018, pp. 1–7, doi: 10.1145/3284557.3284738.
[7] V. K. Sharma, A. Singh, K. R. Jaya, A. K. Bairwa, and D. K. Srivastava, “Introduction to virtualization in cloud computing,” in
Machine Learning and Optimization Models for Optimization in Cloud, 1st Editio., Chapman and Hall/CRC, 2022,
pp. 1–14.
[8] M. Masdari and M. Zangakani, “Green cloud computing using proactive virtual machine placement: Challenges and issues,” Journal
of Grid Computing, vol. 18, no. 4, pp. 727–759, Dec. 2020, doi: 10.1007/s10723-019-09489-9.
[9] M. C. Silva Filho, C. C. Monteiro, P. R. M. Inácio, and M. M. Freire, “Approaches for optimizing virtual machine placement and
migration in cloud environments: A survey,” Journal of Parallel and Distributed Computing, vol. 111, pp. 222–250, Jan. 2018, doi:
10.1016/j.jpdc.2017.08.010.
[10] H. Xing, J. Zhu, R. Qu, P. Dai, S. Luo, and M. A. Iqbal, “An ACO for energy-efficient and traffic-aware virtual machine placement
in cloud computing,” Swarm and Evolutionary Computation, vol. 68, Feb. 2022, doi: 10.1016/j.swevo.2021.101012.
[11] R. Pushpa and M. Siddappa, “Adaptive hybrid optimization based virtual machine placement in cloud computing,” in 2022 4th
International Conference on Smart Systems and Inventive Technology (ICSSIT), Jan. 2022, pp. 1–9, doi:
10.1109/ICSSIT53264.2022.9716298.
[12] H. Zhuang and B. Esmaeilpour Ghouchani, “Virtual machine placement mechanisms in the cloud environments: a systematic
review,” Kybernetes, vol. 50, no. 2, pp. 333–368, Mar. 2021, doi: 10.1108/K-09-2019-0635.
[13] Y. Huang, H. Xu, H. Gao, X. Ma, and W. Hussain, “SSUR: An approach to optimizing virtual machine allocation strategy based
on user requirements for cloud data center,” IEEE Transactions on Green Communications and Networking, vol. 5, no. 2,
pp. 670–681, Jun. 2021, doi: 10.1109/TGCN.2021.3067374.
[14] H. A. Alharbi, T. E. H. Elgorashi, and J. M. H. Elmirghani, “Energy efficient virtual machines placement over cloud-fog network
architecture,” IEEE Access, vol. 8, pp. 94697–94718, 2020, doi: 10.1109/ACCESS.2020.2995393.
[15] Y. Han, D. Guo, W. Cai, X. Wang, and V. C. M. Leung, “Virtual machine placement optimization in mobile cloud gaming through
QoE-oriented resource competition,” IEEE Transactions on Cloud Computing, vol. 10, no. 3, pp. 2204–2218, Jul. 2022, doi:
10.1109/TCC.2020.3002023.
[16] M. Sharma, M. Kumar, and J. K. Samriya, “An optimistic approach for task scheduling in cloud computing,” International Journal
of Information Technology, vol. 14, no. 6, pp. 2951–2961, Oct. 2022, doi: 10.1007/s41870-022-01045-1.
[17] K. Braiki and H. Youssef, “Multi-objective virtual machine placement algorithm based on particle swarm optimization,” in 2018
14th International Wireless Communications & Mobile Computing Conference (IWCMC), Jun. 2018, pp. 279–284, doi:
10.1109/IWCMC.2018.8450527.
[18] M. Kumar, S. C. Sharma, A. Goel, and S. P. Singh, “A comprehensive survey for scheduling techniques in cloud computing,”
Journal of Network and Computer Applications, vol. 143, pp. 1–33, Oct. 2019, doi: 10.1016/j.jnca.2019.06.006.
[19] N. Chaurasia, M. Kumar, R. Chaudhry, and O. P. Verma, “Comprehensive survey on energy-aware server consolidation techniques
in cloud computing,” The Journal of Supercomputing, vol. 77, no. 10, pp. 11682–11737, Oct. 2021, doi: 10.1007/s11227-021-
03760-1.
[20] M. A. Kaaouache and S. Bouamama, “An energy-efficient VM placement method for cloud data centers using a hybrid genetic
algorithm,” Journal of Systems and Information Technology, vol. 20, no. 4, pp. 430–445, 2018, doi: 10.1108/JSIT-10-2017-
0089.
[21] V. Dinesh Reddy, G. R. Gangadharan, and G. S. V. R. K. Rao, “Energy-aware virtual machine allocation and selection in cloud
data centers,” Soft Computing, vol. 23, no. 6, pp. 1917–1932, Mar. 2019, doi: 10.1007/s00500-017-2905-z.
[22] R. Dhaya et al., “Energy-efficient resource allocation and migration in private cloud data centre,” Wireless Communications and
Mobile Computing, vol. 2022, pp. 1–13, Feb. 2022, doi: 10.1155/2022/3174716.

4751
[23] F. Alharbi, Y.-C. Tian, M. Tang, W.-Z. Zhang, C. Peng, and M. Fei, “An ant colony system for energy-efficient dynamic virtual
machine placement in data centers,” Expert Systems with Applications, vol. 120, pp. 228–238, Apr. 2019, doi:
10.1016/j.eswa.2018.11.029.
[24] M. Ranjbari and J. Akbari Torkestani, “A learning automata-based algorithm for energy and SLA efficient consolidation of virtual
machines in cloud data centers,” Journal of Parallel and Distributed Computing, vol. 113, pp. 55–62, Mar. 2018, doi:
10.1016/j.jpdc.2017.10.009.
[25] B. Akay, D. Karaboga, B. Gorkemli, and E. Kaya, “A survey on the artificial bee colony algorithm variants for binary, integer and
mixed integer programming problems,” Applied Soft Computing, vol. 106, Jul. 2021, doi: 10.1016/j.asoc.2021.107351.
[26] S. Hosseini and A. Al Khaled, “A survey on the imperialist competitive algorithm metaheuristic: Implementation in engineering
domain and directions for future research,” Applied Soft Computing, vol. 24, pp. 1078–1094, Nov. 2014, doi:
10.1016/j.asoc.2014.08.024.
[27] B. Abdi, H. Mozafari, A. Ayob, and R. Kohandel, “Imperialist competitive algorithm and its application in optimization of laminated
composite structures,” European Journal of Scientific Research, vol. 55, no. 2, pp. 174–187.
[28] R. Yazdani, M. Hajimahmoodzadeh, and H. R. Fallah, “Adaptive phase aberration correction based on imperialist competitive
algorithm,” Applied Optics, vol. 53, no. 1, pp. 132–140, Jan. 2014, doi: 10.1364/AO.53.000132.
[29] P. Humane and J. N. Varshapriya, “Simulation of cloud infrastructure using CloudSim simulator: A practical approach for
researchers,” in 2015 International Conference on Smart Technologies and Management for Computing, Communication, Controls,
Energy and Materials (ICSTM), May 2015, pp. 207–211, doi: 10.1109/ICSTM.2015.7225415.
[30] E. Rani and H. Kaur, “Study on fundamental usage of CloudSim simulator and algorithms of resource allocation in cloud
computing,” in 2017 8th International Conference on Computing, Communication and Networking Technologies (ICCCNT), Jul.
2017, pp. 1–7, doi: 10.1109/ICCCNT.2017.8203998.
BIOGRAPHIES OF AUTHORS
Seyyed-Mohammad Javadi-Moghaddam was born in Qaen, Iran, in 1975. He
received the B.E. degree in computer engineering from the Ferdowsi University of Mashhad,
Iran, in 1998, the MSc degree in Computer Software from AZAD University of Mashhad, Iran,
in 2007, and Ph.D. degree in Computer Science from National Technical University of Athens,
Greece in 2016. In 2007, he joined the Department of Computer Engineering, PNU University
of Iran, as a Lecturer. Since September 2016, he is with the Department of Computer
Engineering, Bozorgmahr University of Qaenat, Iran as an Assistant Professor. His current
research interests include cloud computing, imbalanced data, and distributed systems. He can be
contacted at email: smjavadim@buqaen.ac.ir.
Zahra Dehghani was born in Qaen, Iran, in 1985. She received a B.E. degree in
computer engineering from the Payam Noor University of Birjand, Iran, in 2007. She took a
Master’s degree in computer engineering from the AZAD University of Birjand, Iran, in 2020.
She can be contacted at email: dehghanizahra90@gmail.com.

Virtual machine placement in cloud using artificial bee colony and imperialist competitive algorithm

More Related Content

Similar to Virtual machine placement in cloud using artificial bee colony and imperialist competitive algorithm (20)

More from IJECEIAES (20)

Recently uploaded (20)

Virtual machine placement in cloud using artificial bee colony and imperialist competitive algorithm