Journal of Physics: Conference Series

PAPER • OPEN ACCESS You may also like

- The mutual influence approach during the
Research on resource allocation strategy of data resource allocation process in construction
projects
center based on cloud computing Nbras Dayoub and Muhammet Fakhratov

- Networking—a statistical physics

perspective
To cite this article: Shanshan Ji 2021 J. Phys.: Conf. Ser. 2083 042085 Chi Ho Yeung and David Saad

- Relationship between the digital economy,

resource allocation and corporate carbon
emission intensity: new evidence from
listed Chinese companies
View the article online for updates and enhancements. Pengyu Chen

This content was downloaded from IP address 14.139.228.226 on 28/04/2023 at 12:46

2021 2nd International Conference on Applied Physics and Computing (ICAPC 2021 IOP Publishing
Journal of Physics: Conference Series 2083 (2021) 042085 doi:10.1088/1742-6596/2083/4/042085

Research on resource allocation strategy of data center based

on cloud computing

Shanshan Ji*
School of Tianjin University of Technology, Tianjin, China

*Corresponding author e-mail: [email protected]

Abstract. With the development of Internet and information technology, cloud

computing has attracted extensive attention from industry and academia. The large scale
of resources, concurrent execution of multiple tasks and dynamic changes of application
resource requests make the resource allocation of data center face severe challenges. To
solve the problem of low balance of traditional resource allocation, this paper focuses
on the resource allocation optimization of data center, and proposes the resource
allocation strategy of data center based on cloud computing, so as to complete the
effective resource allocation and assignment. This paper also verifies the designed
resource allocation method through example research. The research shows that the
distribution balance degree of resource allocation strategy based on cloud computing is
significantly higher than the control group, which proves that the designed resource
allocation strategy can solve the problem of low balance of traditional resource
allocation.

1. Introduction
With the increasing popularity of Internet technology, Internet of Things technology, 5G communication
technology and intelligent terminal equipment, diversified data shows geometric growth, and the
resource allocation of data center has been paid attention to and concerned in many aspects. As a central
organization for data statistics, analysis and processing, the data center realizes the organic integration
of a mass of data resources through a special network virtual dedicated line. The data center is the basis
of cloud computing services. People's requirements for data center are increasing, and the users of data
center are also gradually increasing. Due to the different load of data center resources in different periods
of time, the uneven resource allocation of data center often occurs. Facing many users, the data center
has too much load, large data resource base and complex types, which brings new problems and
challenges to the resource allocation of data center [1], and the resource allocation strategy is an
important means to realize the balanced resource allocation.
Since the concept of cloud computing was put forward in 2006, it has led the third technological
revolution of the Internet. Cloud computing has become the focus of the transformation and strategic
layout of major information technology enterprises around the world. Compared with the traditional
resource allocation, the resource allocation strategy based on cloud computing needs to consider the
problem of balance. Because the traditional resource allocation method of data center has the
phenomenon of low balance, they can’t realize balanced resource allocation, resulting in low resource
allocation efficiency. Therefore, it is necessary to optimize the resource allocation strategy to meet the

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
2021 2nd International Conference on Applied Physics and Computing (ICAPC 2021 IOP Publishing
Journal of Physics: Conference Series 2083 (2021) 042085 doi:10.1088/1742-6596/2083/4/042085

personalized needs of users in many ways. In the past, data center resources focused on the input quantity
of users, and there were still some deficiencies in the balanced allocation of resource channels for clients.
The resource allocation strategy of data center based on cloud computing can realize load balancing
through distributed computing, reasonably allocate resources, and then improve the balance degree. To
solve the problem of resource allocation as soon as possible and make the resource allocation of data
center more balanced and reasonable, this paper designs the resource allocation strategy of data center
based on cloud computing, which is committed to greatly improving the resource allocation balance of
data center through cloud computing.

2. Cloud Computing
Cloud computing not only represents the innovation and optimization of science and technology, but
also represents a new concept and service method. In the cloud environment, cloud users don’t need to
be bound and restricted by time and place. They can use terminal devices anytime and anywhere to
directly obtain the information and resources they want to query from cloud services through the cloud
network [2]. Through cloud computing, data processing can be completed in only a few seconds. The
application of cloud computing in the data center can effectively improve the computing and analysis
ability of data center, which is of great significance to the resource allocation of data center.
Cloud computing technology is a data technology means widely used in various fields. At the same
time, it is also a computing platform environment characterized by distributed structure developed in
recent years. In cloud computing technology, data center is not only the fundamental support and core,
but also the main location for centralized data storage and management. Cloud computing is
characterized by low cost and high efficiency. Generally, cloud computing platform includes four basic
hierarchies: upper computer display layer, middle layer, analysis and processing layer and server
terminal layer. Through the practical application of cloud computing, it can effectively improve the
interconnection between various data, information and other resources and host computer, and realize
the efficient utilization of various virtual resources [3]. How to effectively allocate resources has become
an important problem to be solved in cloud computing, and it is also a hot topic in cloud computing
research.

3. Resource allocation problem

As the data center is becoming large-scale and globalized, it is difficult to allocate and manage data
center resources. On the one hand, the resource consumption of data center increases sharply, and the
resource utilization rate is generally low, resulting in high operating costs of data center [4]; On the
other hand, the resource allocation of data center is often unbalanced. The unbalanced load leads to
excessive virtual machine migration, which consumes the network resources, and the unbalanced use of
resources in the server leads to the waste of resources [5].

4. Resource allocation model construction based on cloud computing

4.1. Resource allocation objective function

In this paper, the first step of constructing the resource allocation strategy of data center is to propose
the objective function of resource allocation. Considering that in the data center, each node performs
different resource allocation tasks and cannot achieve balanced allocation, and some network nodes are
allocated too many resource tasks, which will lead to excessive node energy consumption and failure of
allocation tasks or node functions [6], in order to effectively prevent this phenomenon, a distribution of
node delay minimum objective function is:
H

Z = � Bi + � TSij
i=1

Where i is the number of resource allocation; H is the resource set of data center; T is resource
allocation time of data center; S is the signal-to-noise ratio of data center resource receiver.

2
2021 2nd International Conference on Applied Physics and Computing (ICAPC 2021 IOP Publishing
Journal of Physics: Conference Series 2083 (2021) 042085 doi:10.1088/1742-6596/2083/4/042085

4.2. Resource allocation parameters

According to the objective function of resource allocation in the data center, since cloud computing
requires few estimated parameters, this paper makes a maximum posteriori hypothesis based on the load
balancing assumption of cloud computing to make the resource attributes independent of each other. Set
the maximum posteriori hypothesis and objective function of resource allocation attributes as follows:
𝑛𝑛
𝑝𝑝(𝑐𝑐𝑖𝑖 )
M= � 𝑝𝑝�𝑥𝑥𝑗𝑗 �𝑐𝑐𝑖𝑖 � 𝑍𝑍
𝑝𝑝(𝑥𝑥)
𝑗𝑗=1

N = ηLn + B = � �ηlogv p�v�n(v)��

v
Where p is the number of servers in the data center; 𝑐𝑐𝑖𝑖 is the maximum resource capacity of servers
under the resource allocation times of data center; x is the class with the highest posteriori probability
of resource allocation; n is the amount of resources wasted during resource allocation; j is the number
of virtual machines; Z is the minimum objective function; η is a parameter of resource allocation
structure characteristics; B is the intermediate vector; v is the auxiliary vector. The resource allocation
parameters are calibrated based on the representation vector with high discrimination, and the resource
allocation balance degree is improved to a certain extent by integrating the allocation balance factors
into the objective function of the representation vector [7].

4.3. Resource allocation fitness ranking

After the calibration of resource allocation parameters, the fitness of the dynamic resource allocation is
calculated, and the dynamic allocation of data resources with different priorities is realized by ordering
the fitness of dynamic resource allocation in the data center. Cloud computing is used to divide the
fitness level of dynamic resource allocation, so as to avoid conflicts and low priority queue hunger in
the process of dynamic resource allocation. Suppose that calculation expression of the fitness of dynamic
allocation of data center resources is:
i
W=
∑ni=1 i
According to the above formula, the fitness of dynamic resource allocation of data center can be
obtained. On this basis, the data center resource dynamic allocation fitness ranking is carried out based
on the cloud computing allocation fitness ranking mechanism. Combined with the sorting mechanism,
the fitness of dynamic resource allocation of data center is used as the selection operator in this allocation.
While preserving the diversity of dynamic allocation of data center resources by means of mutation
evolution, the similarity between dynamic allocation of resources is distinguished by sorting the fitness
of dynamic allocation of data center resources.

4.4. Resource allocation implementation

Build a complete resource allocation method of data center based on cloud computing through the fitness
ranking of dynamic allocation of data center resources, find the optimal solution based on cloud
computing, and realize the balanced resource allocation of data center [8], as shown in Figure 1.

3
2021 2nd International Conference on Applied Physics and Computing (ICAPC 2021 IOP Publishing
Journal of Physics: Conference Series 2083 (2021) 042085 doi:10.1088/1742-6596/2083/4/042085

Figure 1. Flow chart of resource allocation.

The objective function ω = (𝑃𝑃𝑚𝑚𝑚𝑚𝑚𝑚 , A, O, Z, l, f, D, P𝑚𝑚𝑚𝑚𝑚𝑚 ) represents the resource allocation of data
center, where O is the resource allocation paths; l is the average connectivity of resource allocation
network; f is the communication resource quantity on the resource allocation paths; D is the
congestion degree of data center nodes. Based on the objective function of resource allocation based on
cloud computing, through the proportional fairness algorithm to ensure the balance of resource
allocation, sets the expression for the boundary function ∂, and obtain the complete resource allocation
expression based on cloud computing:
∂ = e1 p1 (t) + e2 p2 (t) + ⋯ + ek pk (t)
m. t max(ω∗ )
⎧
s. t1 maxE(B)
ρi =
⎨ s. t2 maxE(γ)
⎩s. t3 P(ω) > {(1 − u1 )u2 − s}/ ∂
Where 𝑒𝑒1 is resource allocation transmission power; 𝑝𝑝1 is the average speed of each data center
resource combination; 𝑡𝑡 is the maximum cellular signal-to-noise ratio; 𝑘𝑘 is the number of resource
nodes; E is the characteristic of resource allocation; γ is the analytical parameter of resource
allocation behavior; 𝑢𝑢1 is the probability of high quality resource allocation; 𝑢𝑢2 is the probability of
low quality resource allocation. The optimal solution with the maximum value of 𝜌𝜌𝑖𝑖 is to perform
resource allocation.

5. Experimental study

5.1. Experimental design

In order to verify the rationality and practicability of the designed resource allocation strategy, through
the method of experimental study, this paper takes a data center as the experimental object, we randomly
selects two real data sets (Tyhusge data set and Cfisnja data set) as the data center resources, selects a
circular area with a side length of 10 km, and sets 200 communication nodes. The distribution balance
degree of resource allocation by cloud computing method and traditional method is tested by Kerterly
software. The former is recorded as the experimental group and the latter as the control group. Set the

4
2021 2nd International Conference on Applied Physics and Computing (ICAPC 2021 IOP Publishing
Journal of Physics: Conference Series 2083 (2021) 042085 doi:10.1088/1742-6596/2083/4/042085

allocation times to 10 times, and take every 2 times as a recording node to record the experimental
results. The higher the degree of distribution balance is, the higher the rationality of the distribution
strategy is.

5.2. Experimental results

According to Table 1, in the 10th experiment, the highest value of distribution balance degree in the
experimental group is 1.57, in the 4th experiment, the highest value of distribution balance degree in the
control group is 0.78, and the experimental group is 0.79 higher than the control group; In the 2th
experiment, the lowest value of distribution balance degree in the experimental group is 1.31. In the 8th
experiment, the lowest value of distribution balance degree in the control group is 0.55, and the
experimental group is 0.76 higher than the control group. Therefore, the distribution balance degree of
the resource allocation strategy designed in this paper is higher than the control group, which can realize
the balanced distribution of data center resources and has practical application value.

Table 1. Distribution balance degree table.

frequency experimental group control group
2 1.31 0.63
4 1.40 0.78
6 1.36 0.60
8 1.35 0.55
10 1.57 0.65

6. Conclusion
This research makes cloud computing from theory to practice. Based on cloud computing, load
balancing and resource allocation can be solved. It is no longer a simple concept at first. Various
technologies of cloud computing have made breakthrough progress. This paper designs the resource
allocation strategy of data center based on cloud computing, constructs the objective function, verifies
the practicability and rationality of the designed resource allocation strategy by means of example
analysis, and proves that it can solve the defects existing in the traditional resource allocation, prevent
the waste of resources caused by unreasonable load allocation, and realize the balanced allocation of
data center resources.
Although the experimental analysis is carried out in this paper, there are still some shortcomings.
The accuracy of the experimental analysis results is not tested. In the future, we can supplement the
research in this regard, deeply study the resource allocation and optimization of data center, and provide
constructive suggestions to improve the quality of resource allocation.

References
[1] Luo, P.& X. Wang.Research on the resource allocation algorithm based on forecasting in mobile
cloud computing [J]. Journal of Intelligent & Fuzzy Systems, 2018, 35(2): 1315-1324.
[2] Geelan, J.Twenty-one experts define cloud computing [J]. Cloud Computing Journal, 2009(4): 1-
5.
[3] Mala, K.& S. Sarbjeet.A review of metaheuristic scheduling techniques in cloud
computing[J].Egyptian Informatics Journal, 2015, 16(3): 275-295.
[4] Al-Dulaimy, A., W. Itani.,Zekri A, et al.Power management in virtualized data centers:state of
the art [J]. Journal of Cloud Computing, 2016, 5(1): 6.
[5] Wei, W., Y Liu.& W Yang.A fast approximation algorithm for the general resource placement
problem in cloud computing platform [J]. Journal of Computer Reserch and Development,
2017, 9(3): 17-33.
[6] Gawali, M.B.& S.K. Shinde.Task scheduling and resource allocation in cloud computing using a
heuristic approach [J]. Journal of Cloud Computing, 2018, 7(1): 1-16.
[7] Vicentini, C., S. Altair, V. Eduardo, et al.SDN-based and multitenant-aware resource

5
2021 2nd International Conference on Applied Physics and Computing (ICAPC 2021 IOP Publishing
Journal of Physics: Conference Series 2083 (2021) 042085 doi:10.1088/1742-6596/2083/4/042085

provisioning mechanism for cloud-based big data streaming [J]. Journal of Network and
Computer Applications, 2018.
[8] Cheng, M.Y.& D. Prayogo.Symbiotic Organisms Search:A new metaheuristic optimization
algorithm [J]. Pergamon, 2014(139): 98-112.