Improvement of the Secure Integration of IoT and Cloud Computing
using Hybrid Encryption
P. DR. NADEEM CHAHIN 1, Eng. ABEER MANSOUR 2
Electronics and Communication Engineering 1, 2
Damascus University 1, 2
DAMASCUS, SYRIA
Abstract: Wireless Sensor Network (WSN) is an essential technology in many Internet of Things (IoT)
applications, and since sensor nodes suffer from limited resources, it has become possible to overcome storage
capacity problems using cloud computing, the integration of Internet of Things (IoT) with cloud computing (CC)
seeks to achieve new levels of efficiency in service delivery. Security and privacy are key factors that slow down
the rapid and widespread adoption and deployment of both IoT and cloud computing. In the proposed model, an
integrated IoT system with cloud computing was developed starting from the analysis, and design, to the
implementation to connect IoT devices with the cloud, the security is achieved by using a hybrid encryption
mechanism which provides the performance advantages of symmetric and asymmetric encryption algorithms.
Where the Elliptic Curve Cryptography (ECC) algorithm is used for key generation and AES (Advanced
Encryption Standard) algorithm is used for encryption and decryption of the sensors' data to provide a reliable
computing environment. We have implemented the proposed system and showed the results of using CONTIKI
COOJA 3.0 that connected with the cloud service provider, Evaluate a set of performance metrics such as power
consumption, packet delivery ratio, and the algorithm execution time, in addition to verifying network immunity
against the black hole attack.
Key Words: Internet of Things, Cloud Computing, Security, Hybrid Encryption, Reliability.
Received: October 22, 2022. Revised: November 15, 2022. Accepted: November 30, 2022. Published: December 8, 2022.
1 Introduction
From their emergence, the two concepts of the
Internet of Things (IoT) and cloud computing (CC)
have evolved separately. For many years, they have
seen independent evolution in their hardware and
software aspects. In its evolution, IoT faces many
problems among them storage capacity, energy
efficiency, and computational capabilities. While
looking for solutions to these problems, scientists
found that CC could help to solve them. In addition,
the Internet of Things could allow CC to handle real-
world objects in a more dynamic way to deliver new
attractive services and applications in some practical
applications.
Hence, the need to merge the Cloud and IoT
technologies emerge. As a result, the concept of the
Cloud of Things (CoT) was born. This integration is
useful because the resulting system is more powerful,
and intelligent and offers promising solutions to the
users. However, CoT faces a large number of
challenges such as security, privacy, reliability,
scalability, heterogeneity, power consumption,
standardization, and others [1].
Security and privacy are frequently discussed
when talking about IoT and cloud computing
integration. since the cloud consists of a large number
of servers that host applications responsible for
carrying out computer processing operations on data
collected from IoT devices[2], the network will be
vulnerable to many attacks such as MIM (Man in the
Middle), Sinkhole attack, Vampire Attack, Jamming
Attack. One of the most important attacks that limited
resources networks suffer from is the blackhole
attack, in which the attacking node drops all packets
sent to another node in its root, and aims to steal the
important information or waste the network resources
[3]. To protect the data that is exchanged in the
network and stored in the cloud, encryption
technologies (like lightweight cryptographic
algorithms) are used, which help to securely transmit
data over the wireless medium and provide
International Journal of Electrical Engineering and Computer Science
DOI: 10.37394/232027.2022.4.10
P. Dr. Nadeem Chahin, Eng. Abeer Mansour
E-ISSN: 2769-2507
66
Volume 4, 2022
reliability, confidentiality, data integrity, and non-
repudiation [4].
Within the framework of security and privacy
requirements, and to deal with the challenges of IoT
devices, a lightweight encryption solution based on a
hybrid encryption system is used to provide better
protection for data stored on the cloud. hybrid
encryption is a combination of symmetric and
asymmetric encryption algorithms to provide more
security for the data before it is stored on the cloud.
The remaining part of this paper is organized as
follows. Section 2 provides the related works on the
integration between the internet of things and cloud
computing. Section 3 provides the research materials
and methods. Section 4 shows the performance
results with detailed descriptions. Section 5
concludes the paper.
2 Related Works
The integration between the wireless sensor
networks and cloud computing can be achieved by
creating virtual backups of sensors and their
information on the cloud and the same calculations
are performed on physical and virtual sensors [7]
through a simple model of the principle of
Distributed Shared Memory (DSM).
A different approach in [8] discusses a novel
three-level architecture that integrates WSN, IoT, and
cloud computing capabilities. At the bottom, the SPL
(Sensor Processing Level) clusters all sensors. The
DAL (Data Analysis Level) is considered a local
interface between sensors, the cloud platform, and
the users or health specialists. The DAL server
backend consists of local algorithms for data
aggregation and event-triggered alarms. The graphic
user interface available at this level allows real-time
monitoring of data gathered from sensors. The CPL
(Cloud Prediction Level), allows data storage, data
correlation, and prediction. It generates useful
information for users and specialists. Using this
layer, the health specialist can provide an anticipative
reaction to health issues considering also the air
quality level. The paper provides an implementation
solution for advanced AAL (Ambient Assisted
Living) application implementation. The novelty is
not only in the different processing levels but also in
the implementation that provides solutions for
ensuring a fast reaction speed in case of emergency,
together with access to complex prediction
algorithms.
The research in [9] proposed a general
architecture called (IoTaaS) to represent and guide
the integration process. This architecture supports
security by providing a secure gateway that connects
the IoT system to the cloud, supporting data
encryption for data passed from/to the cloud.
Providing the authentication manager to ensure only
authorized parties have access to the IoT system and
cloud services. Delegating all database operations to
the data manager and handler to ensure only valid
operations are carried out. The architecture also
supports real-time interactions by providing the real-
time interactive handler, which uses artificial
intelligence and context awareness to provide real-
time access to IoT systems. Real-time interaction is
also supported by making modifications to the data
management system and using context information to
reduce response time.
IoT devices are mostly powered by batteries and
have a limited amount of memory and processing
capacity. Designing a security mechanism that can
suit these small devices is a challenging task, and
there are concerns about cloud administrators who
can access the collected data. and to protect this data,
various technologies such as encryption have been
proposed [2].
We can classify the encryption algorithms into
two general categories: the Symmetric Encryption
Algorithm (Private-Key) and the Asymmetric
Encryption Algorithm (Public-Key) [4] shown in
Fig.1. The symmetric key technology uses a single
key called the secret key that uses less mathematics,
and less calculation, on the other hand, the
asymmetric key technology uses both public and
private keys, leads to more processing and consumes
more power but the keys are more strong and
secure[6].
Fig.1 Classification of cryptographic algorithms [6]
The results of LabVIEW 2016 simulations to
study several encryption algorithms in [10] showed
that AES is the best encryption algorithm in terms of
speed, productivity, and low power consumption
because in the 128-bit key encryption it can encrypt
a 128-bit block (the largest encryption block for
encryption algorithms) in just 14 cycles. The DES
(Data Encryption Standard) algorithm then comes in
second place as the best performance, while 3DES
International Journal of Electrical Engineering and Computer Science
DOI: 10.37394/232027.2022.4.10
P. Dr. Nadeem Chahin, Eng. Abeer Mansour
E-ISSN: 2769-2507
67
Volume 4, 2022
(Triple DES) and RC2 have the same level of
performance, RSA (Rivest–Shamir–Adleman) has
the lowest level of performance in terms of speed and
productivity.
In the research [11] a hybrid encryption system is
proposed using different encryption algorithms to
provide the advantages of both symmetric and
asymmetric encryption using ECDH (Elliptic-curve
Diffie–Hellman) to generate keys and perform a
digital signature to achieve authentication, and AES
is used for encrypting the data.
The researchers in [13] proposed a hybrid
encryption approach combining the ECC algorithm
and the ElGamal algorithm. The encrypted
information is uploaded to the cloud for storage, in
different channels and passwords. The user who
authenticates the strategies can download the
encrypted data and decrypt them by a synchronous
key that uses the ECC-ElGamal algorithm. The
simulation was performed using MATLAB 2014a
and calculating various performance factors such as
execution time, packet delivery rate, and delay, and
comparing these results with the traditional methods
and found that it offers 12%, 31%, and 8% in terms
of packet delivery rate, delay and execution time in
respect.
3 Research materials and methods
The work was performed using a computer with a
core i5 processor (2.20 GHz) – 8 GB RAM Windows
10 operating system (64 Byte). virtual tool (VMware
virtual machine) was used to run Ubuntu (64 Byte
and 8 GB RAM) and the simulation was performed
on CONTIKI COOJA 3.0 which supports the Internet
of Things networks and all protocols in the field of
networks. It is considered an emulator and not a
simulator because its performance is closer to reality
and it performs a real operation on the sensors
devices in the network. The work has been done on
Skymote nodes type, the simulation language used is
C.
In the proposed model, the network was implemented
according to the following steps, fig.2:
1. Choose the gateway node that connects the nodes
to the Internet.
2. Discover the nodes by the sink node.
3. Generate keys by sink node using the ECC
algorithm and distribute them to nodes in the
network.
4. Encrypt the sensor data using the AES algorithm
at the nodes.
5. Transfer the encrypted data from nodes to the
cloud server.
6. Study a set of metrics such as power consumption,
execution time, and packet delivery ratio, and
compare the results with a reference model and with
other studies.
7. Apply and detect a blackhole attack on the
network to verify the immunity of the network and
the effectiveness of the proposed mechanism.
Fig.2 Flowchart of the proposed model
The proposed model was achieved in three scenarios
on an increasing number of nodes (10-15-20 nodes in
order) with a network area (100*100 m2):
Scenario I (no encryption): a study of the
Internet of Things system connected to the
cloud server without applying any
encryption algorithms.
Scenario II (encryption): apply the proposed
hybrid encryption algorithm on the same
network in the first scenario and compare the
results for both packet delivery ratio, power
consumption, and execution time.
Scenario III (blackhole attack): Apply the
blackhole attack on the network and detect
the attack while explaining the usefulness of
the proposed method of protection.
The COOJA is connected to a cloud using
smarterasp.net which gives the possibility of building
a private cloud from the beginning, downloading and
modifying algorithms as needed, fig.3.
Fig.3 Cloud Service Program Interface
International Journal of Electrical Engineering and Computer Science
DOI: 10.37394/232027.2022.4.10
P. Dr. Nadeem Chahin, Eng. Abeer Mansour
E-ISSN: 2769-2507
68
Volume 4, 2022
The data collected from the sensors is sent in
encrypted form to the cloud server for storage, the
decryption can be performed only by users
authorized to enter and had the decryption ECC code
of the network which ensures confidentiality and
reliability.
Fig.4 Displaying encrypted sensors data on the
cloud site
View sensors data after decryption using a special
decryption code:
Fig.5 Display the sensor data after decoding on the
cloud site.
4 Results and comparison
The performance of the proposed model was tested,
analyzed, and compared with the reference state and
then the results were compared with existing studies.
4-1 Packet Delivery Ratio (PDR %):
PDR (Packet Delivery Ratio) refers to the ratio of the
total number of packets received at the destination to
the total number of packets created during the
simulation, expressed according to the equation [5]:
Where S is a set of sessions created during the
simulation and npacketsReceived (i.d) and
npacketsSent (i.s) are the number of packets received
at the destination d and sent from source s for session
i, respectively.
We note that for the first and second scenarios there
is no loss of packets where the delivery ratio is 100%.
The proposed algorithm did not cause any loss of
packets in the network. At the attack scenario, we can
find that the attacking node causes a significant loss
of packets, in fig.6.
The attack can be detected by monitoring the number
of received packets because the noticeable increase
in the total number of messages exchanged in the
network nodes is due to the increase in the number of
control messages sent by the victim nodes trying to
rejoin the network, shown in fig.7.
Fig.6 Packet Delivery Ratio%
Fig.7 Number of total packets in case of attack
4-2 Power Consumption:
The simulation gives a set of values related to power
consumption by each node individually and in detail
(CPU power, LPM power, Listen power, Transmit
power) and the result of power consumption for each
node, which is the sum of all previous values.
Expressed according to the equation [14]:
Energestvalue , Current, and Voltage are taken from the
simulator and RTIMER_SECOND is a Clock.
International Journal of Electrical Engineering and Computer Science
DOI: 10.37394/232027.2022.4.10
P. Dr. Nadeem Chahin, Eng. Abeer Mansour
E-ISSN: 2769-2507
69
Volume 4, 2022
The small increase in the power consumed due to the
encryption process is acceptable compared to other
studies and does not affect the functioning of the
network.
When applying the blackhole attack, a significant
increase in power consumption can be observed, and
the increase is caused by the desire of the victim
nodes to rejoin the network and send more control
messages, Fig.8.
Fig.8 Power Consumption (mW)
4-3 Execution time of the algorithm:
The Execution time of the algorithm was obtained by
using a computer Core i5 processor (7th generation)
and (8RAM) and SSD hard because the results in the
encryption speed change depending on the state of
the device, all values measured by ms, the keys are
distributed after detection of all the surrounding
nodes, with a maximum simulation runtime (2.7 m)
for the 10-node scenario. Fig.9 shows in detail the
time required for the process of key generation,
encryption, and decryption, and the total execution
time of the algorithm.
Fig.9 Algorithm execution time (ms)
4-4 compare the results of the proposed model
with existing studies:
Compare power consumption:
The researchers proposed a hybrid encryption model
in [14] using the AES algorithm to encrypt data with
ECDH to generate keys and authentication, and we
note that the proposed model offers suitable and
better solutions for energy consumption as the data
exchange rate increases in the network,fig.10.
Fig.10 Comparison of power consumption with [14]
Comparison of algorithm execution time:
Comparing the study with [12], we find that the
proposed model is reducing the execution time of the
algorithm that results in less delay and lower power
consumption, so we can find that the proposed model
offers a lightweight and fast security solution, fig.11.
Fig.11 Comparison of the execution time of the
algorithm with other studies [12-14]
4-5 Detection of the blackhole attack:
The attacking node aims to disrupt the service and
steal the important information of the users. In the
proposed model, the attack can be detected when a
new node tries to join the network and obtain
authentication, which will cause high-energy
consumption, as in Fig.8. Also, the increasing
number of packets with an attempt to join the network
by an attacking node from Fig.7. When the attacking
node is detected, it will be prevented from joining the
International Journal of Electrical Engineering and Computer Science
DOI: 10.37394/232027.2022.4.10
P. Dr. Nadeem Chahin, Eng. Abeer Mansour
E-ISSN: 2769-2507
70
Volume 4, 2022
network, and encryption keys will be redistributed
immediately from the sink node.
In addition, the attacking node will not be able to
decrypt the data in the network due to the encryption
process and the strength of the keys used.
5 Conclusions and recommendations
The model proposed in this paper helps to achieve
secure integration between the Internet of Things and
cloud computing using the hybrid encryption model
(AES-ECC). To secure a reliable connection between
the sensor nodes in the network and the cloud, store
the sensor data securely, encrypted without affecting
the packets delivery ratio, power consumption, and
provide lower execution time compared to the others
studies. The combined effect of both ECC and AES
is appropriate for the proposed cloud storage
technology to secure the system and helps to reduce
storage volume with encrypted data.
This research aims to combine the availability of
resources and maintain reliability and privacy when
achieving the integration between the Internet of
Things and cloud computing, and since each
application has its requirements and conditions so
this study is mainly directed to the Internet of Things
networks with fixed and random publishing nodes in
a field of study where data is collected from sensors
and sent to the cloud to store them confidentially,
Evaluate a set of parameters such as average power
consumption, packet delivery ratio, and the algorithm
execution time and compared to previous studies.
The model can be used in many IoT applications that
need speed, limited resources, and confidentiality in
the transmission and storage of sensitive data.
However, the concept of security needs more study
in the future to expand the concept of cloud
computing through encryption technologies. In the
future, this research could be improved by increasing
the security of hybrid approaches by adding multiple
layers of security to enhance system productivity and
efficiency.
References:
[1]Ari, A. A. A., Ngangmo, O. K., Titouna, C.,
Thiare, O., Kolyang, Mohamadou, A., Gueroiu, A.
M. (2019), “Enabling privacy and security in Cloud
of Things: Architecture, applications, security &
privacy challenges” Applied Computing and
Informatics, Vol. ahead-of-print No. ahead-of-print,
page 3.
[2]Rayes, A., and Salam, S., 2019, “Internet of
Things from Hype to Reality_ the Road to
Digitization” Second Edition, © Springer Nature
Switzerland AG.
[3]Sokat, B., 2020, “Blackhole Attacks in IoT
Networks" Master Thesis, Computer Engineering,
Engineering and Sciences of İzmir Institute of
Technology, İZMİR, 43p.
[4]Salman, O., Abdallah, S., Elhajj, I. H., Chehab,
A., and Kayssi, A., 2016, “Identity-Based
Authentication Scheme for the Internet of Things”,
978-1-5090-0679-3/16/$31.00 ©2016 IEEE.
[5]Ahmed, F., and Young-Bae Ko, 2016,
“Mitigation of black hole attacks in Routing
Protocol for Low Power and Lossy Networks”,
Wiley Online Library (wileyonlinelibrary.com).
DOI: 10.1002/sec.1684
[6]HASSAN, A., ISMAIL, A., 2017, “Evaluation of
encryption algorithms for IoT security” Bachelor of
Science in Communication Engineering,
Communication Engineering Department,
University of almughtaribeen, Khartoum, Sudan,
49p.
[7]Langendoerfer, P., Piotrowski, K., Díaz, M. ,and
Rubio, B., 2012, “Distributed Shared Memory as an
Approach for Integrating WSNs and Cloud
Computing” IEEE.
[8]Chenaru, O., Mihai, V., Popescu, D., and Ichim,
L., 2018, Member, IEEE “Integration of WSN, IoT
and Cloud Computing in Distributed Monitoring
System for Aging Persons in Active Life” 26th
Mediterranean Conference on Control and
Automation (MED) Zadar, Croatia.
[9]Othman, M. M., El-Mousa, A., 2020, “Internet
of Things & Cloud Computing Internet of Things as
a Service Approach”, 2020 11th International
Conference on Information and Communication
Systems (ICICS), Amman, Jordan.
[10]Latif, I.H., 2020, “Time Evaluation Of
Different Cryptography Algorithms Using
Labview” 2020 IOP Conf. Ser.: Mater. Sci. Eng.
745 012039View
International Journal of Electrical Engineering and Computer Science
DOI: 10.37394/232027.2022.4.10
P. Dr. Nadeem Chahin, Eng. Abeer Mansour
E-ISSN: 2769-2507
71
Volume 4, 2022
[11]Kunchok, T., Prof. Kirubanand V.B., 2018 “A
Lightweight hybrid encryption technique to secure
IoT data transmission”International Journal of
Engineering & Technology, 7 (2.6) 236-240.
[12]Rehman, S., Bajwa, N.,Shah, M., Aseeri , A.,
and Anjum, A.,2021, “Hybrid AES-ECC Model for
the Security of Data over Cloud Storage”
Electronics 2021, 10, 2673.
[13]Kumari C. S., Asha C. N., Rajashekhar U., K.
Viswanath, 2022, “Performance Analysis of Cloud-
based Health Care Data Privacy System Using
Hybrid Techniques”, International Journal Of
Biology And Biomedical Engineering, Volume 16.
[14] Alzahrani, S.M., 2022, “Secure Authenticated
Key Exchange For Enhancing The Security Of
Routing Protocol For Low-Power And Lossy
Networks”, Master thesis of Science in Cyber
Security, Wright State University.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the Creative
Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en_US
International Journal of Electrical Engineering and Computer Science
DOI: 10.37394/232027.2022.4.10
P. Dr. Nadeem Chahin, Eng. Abeer Mansour
E-ISSN: 2769-2507
72
Volume 4, 2022
