Preventing DoS Attacks in IoT Using AES

Full text

e-ISSN: 2289-8131 Vol. 9 No. 3-11 55
Preventing DoS Attacks in IoT Using AES
Yasir Javed1, 2, Adnan Shahid Khan1, Abdul Qahar1, 3 and Johari Abdullah1
1FCSIT, UNIMAS, Sarawak, Malaysia
2Prince Sultan University, Riyadh, KSA.
3The University of Punjab, Lahore, Pakistan.
yjaved@psu.edu.sa
Abstract—The Internet of Things (IoT) is significant in
today’s development of mobile networks enabling to obtain
information from the environment, devices, and appliances. A
number of applications have been implemented in various kinds
of technologies. IoT has high exposure to security attacks and
threats. There are several requirements in terms of security.
Confidentiality is one of the major concerns in the wireless
network. Integrity and availability are key issues along with the
confidentiality. This research focuses on identifying the attacks
that can occur in IoT. Packet filtering and patches method were
used to secure the network and mitigate mentioned attacks but
these techniques are not capable of achieving security in IoT.
This paper uses Advanced Encryption Standard (AES) to
address these mentioned security issues. Official AES version
uses the standard for secret key encryption. However, several
problems and attacks still occur with the implementation of this
original AES. We modified AES by adding white box and the
doubling of the AES encryption. We also replaced the
Substitute-Byte (S-Box) in the conventional AES with the white
box. The significance of a white box is where the whole AES
cipher decomposed into round functions. While doubling the
process of AES gives difficulty to the attacker or malware to
interrupt the network or system. From the algorithms, our
proposed solutions can control DoS attack on IoT and any other
miniature devices.
Index Terms—DoS Attack; AES; IoT Security;
Confidentiality; White Box.
I. INTRODUCTION
The inventions of smart devices and mobile devices are
growing very rapidly as the development of mobile
computing is also increasing with all of them including
Internet of Things (IoT) as their integral part. IoT are objects
that are uniquely identifiable and have Internet-like structure
virtual representation [1]. The growth of IoT nowadays is
enormous. Modern technologies such as Radio-Frequency
Identification (RFID), sensor networks, short-range wireless
communications, and real-time localization are now
becoming extremely usual, which apply IoT into commercial
use. There are wide range of networks such as Wireless
Sensor Network (WSN), Vehicular Ad-Hoc Network
(VANET), Radio-Frequency Identification (RFID),
Smartphone, and others, are included in constructing the IoT
[2]. There are billions of IoT devices connected today and are
expected to increase in coming years. Because of wide range
of applications, elements in IoT interact via broadcasting
messages, which create the messages’ dissemination
efficient. This make IoT network prone to attacks, where
attacker to interrupt the networks. It will be easier for attacker
to intercept, fabricate or even steal the data in the networks
that might be the high-secrecy or private confidential
information. All these suggested solutions are applicable in
the real time applications. People's requirements for
improved living condition are constantly rising because of the
development of economy and the rise of information-based
society. Building Smart Home based on the advancement
information technology is becoming more important. It is
essential to process and use the immense and decentralized
information. Smart home is the basis component of Intelligent
Residential District [3]. The user can correspond with the
latest security dynamics of the whole family if security
devices, such as infrared detector, smoke sensor, etc. can be
accessed to the IoT network. To access to the network of IoT
and community hospital, user need to have household
medical devices like sphygmomanometer so doctors can
correspond with the patients’ health condition timely and
make treatment. Family business center can complete a series
of tasks, such as shopping and payment so people can stay
indoors to manage their daily life [4]. Thus, the security
requirements are the major concern regarding these problems.
One of the key problems is the leak of personal information
that shows the violation of confidentiality. The integrity issue
come through when there is a stealing of data and identity.
Attacks on integrity can disable the sensing and control the
information. Availability is another great target for the
attacks. This research focuses on Denial of Services (DoS)
attack that is high possibility to occur when it comes to
availability context. DoS attack occurs when the system or
service that is required cannot be accessed. Thus, securing
broadcast communication protocol is conducted into
research. This research proposes several improvement in
AES encryption structure. White box as stated, it decomposes
the whole AES cipher into round functions.
Section II presents the related works, about known AES
mechanism and their modification, Section III presents a
system model Section IV presents the proposed solution.
Section V presents the mathematical analysis of proposed
solution. Section VI presents the conclusion.
II. RELATED WORKS
AES replaces the Data Encryption Standard (DES) in 2001
[4-6]. AES contains ten iterations of four key operations,
which are the SubBytes, ShiftRows, MixColumns and
AddRoundKey [7]. Security and confidentiality have always
been the major concern of cloud consumers since the
beginning. Such concerns are derived from several types of
attacks on several systems that have occurred since these
cloud-based services are implemented. Brute force attack,
also known as Dictionary attack or Hybrid Brute-force attack
is one of the incorporate attacks. According to Whitney [8],
brute-force attack is an attack that uses “Trial and Error”
method by guessing users’ passwords. This method initially

Journal of Telecommunication, Electronic and Computer Engineering
56 e-ISSN: 2289-8131 Vol. 9 No. 3-11
initiated by gathering the fundamental information about the
user such as user’s nickname, pet name, birthday month or
year and vehicle name. Attackers usually launch this method
by combining random and guessable passwords such as
‘1234’ or ‘qwerty’. In 2013, October 3 to be exact, Adobe has
confirmed that it had been a victim of an attack that affected
38 million of active users—which mostly used guessable and
random words as their passwords [9]. This brute force attack
enables attackers to guess users’ passwords using a key
derivation function, known as exhaustive key search. This
attack is a cryptanalytic attack that used to decrypt or crack
any encrypted data.
Jose [10] has proposed a technique called “SecCloud
Protocol Implementation” which encrypted protocol data and
sent it to the cloud servers. This technique initially used RSA
algorithm for encryption, but since it suffered from brute
force attack, another AES algorithm has been introduced to
solve this problem. However, brute force attacks still exist
since the author did not focus on the S-box model.
Meanwhile, Junjie and Feng [11] proposed a solution by
using an expansion key on AES algorithm to improve the
security of the key and maintaining the algorithm efficiency
at the same time. This solution proposed does increase the
security, but it does not increase the key length, which might
be a problem since brute force attacks target on key length.
Al-Haj et.al [12] have chosen as the AES in 2001 after an
open process because of its efficiency, security, implement
ability, flexibility, and performance. There are many
applications using AES since it gives many benefits such as
remote diagnosis contribute by telemedicine applications
access to centralized medical remote-distance learning and
medical archives. They suggested improvised method, which
they provide the encryption keys and initialization vectors
externally so that it can overcome the overhead.
White box as stated, it decomposes the whole AES cipher
into round functions. The output of previous round is the
offset of the input of encoding in the ith round. All the
encodings is produced through both linear and non-linear
mapping. Next, the shiftrows transformation that we
suggested is the simple permutation process, which is circular
byte shift that can be guaranteed. Mix columns also
applicable where the four bytes in each column of a state is
combine with an invertible linear transformation. We
proposed the AES add round key algorithm adapted from
Rijndael algorithm[13]. In this algorithm, the 128-bit
independent key for every round is derived from the original
128-bit encryption key. Our research proposed a doubling
process of AES with two different keys. It will ensure the
attacker not to disrupt the network or system because it is
strengthen by the double encryption of AES. AES uses byte
wise substitution, byte swapping and the XOR operation for
encryption process. The mathematical analysis for this
operation is when 128-bit plaintext block are arranged into
4x4 columns. It called ‘states’ and to generate 128 bits of
cipher text it will undergo 10 rounds. At the same time, the
key expansions also occur. The 128-bit keys will produce
other 128-bit keys for 10 rounds. Ten (10) matrices are made
up from the arrangement of keys in 4x4 matrices.
III. SYSTEM MODEL
Figure 1 represents a system design for our proposed model
where a user would like to upload the data into cloud. For
This a Cloud User will use our extended AES Encryption
standard to store the data. Where the Data owner if want to
see the data has to decrypt the cipher using his known key.
Thus, Cloud user will be able to send the data securely over
the cloud and data owner while using the secret key can see
it.
Figure 1: The implementation of AES on IoT (Cloud Service)
Figure 2 represent a DOS attack, where attacker has
stopped the services to Data owner.
Figure 2: The DoS attack on IoT (Cloud Service)
Internet has become a mission- critical component in
modern business. Having said that, cybersecurity has become
indispensable element in the information system. However,
threats and attacks are inevitable in the cyber world due to
lack of security [5]. The IoT is not an exception to this
problem. One of the problems arose is the unwanted release
of personal information. It is the violation of data
confidentiality [14]. For example, a data leak in the smart
medical center monitoring system will lead to an inadvertent
release of sensitive medical data. Therefore, loss of
confidentiality in passwords and important keys will cause
unwanted system threats. Furthermore, attacks on integrity
can distort the sensing and control information [15]. For
instance, the house controller confused by unauthenticated
system status alerts will predicted a situation mistakenly
which allowed an illicit entry. Availability becomes the
greatest target for the attacks. Many Internet-enabled devices
often configured with default or weak password. For instance,

Preventing DoS Attacks in IoT Using AES
e-ISSN: 2289-8131 Vol. 9 No. 3-11 57
in-car Wi-Fi has the same security vulnerabilities as
traditional Wi-Fi hotspots. However, in-car devices and data
will be at risk without the firewalls present in conjunction
with Wi-Fi installation. Hence, as more and more products
are developed with the capacity to be networked, wirelessly
networked IoT devices with a low operational duty cycle will
flood the network and this can lead to a denial of service to
legitimate users.
IV. PROPOSED SOLUTION
A. Reminiscing the Old Solution
As stated in the problem statement, one of the common
services that facing DoS is the network device level.
Previously, the two common ways in preventing this matter
were patches method and packet filtering. For instance, to
hinder the attacks, it is better for a network to reject broadcast
ping requests to reach the network from outside, or to
configure a firewall that rejects all arriving echo request
packets. Suppose that the Internet Service Provider (ISP) in
Figure 1 owns the router that give the end network internet
access and the ISP knows the legal address size of packets
trying to reach through this router. If a packet is trying to
reach the ISP and act to be from an address that is outside of
this legal address space, thus it is clearly depicted. The
deception attempt can be logged and the packet can be filtered
out.
B. Newly Proposed AES Encryption
However, filtering packets and patches method were not
robust enough to secure the network security. Here, we have
indicated a couple of new algorithm to modify the existing
AES, which are the doubling of the AES encryption, and the
white box, which replaces the Substitute Bytes (S-Box).
Below are the explanation for the method used in AES
encryption that can prevent the attack techniques of DoS such
as Attack Tools, Application-layer Floods, Degradation-Of-
Services and Denial-Of-Services from occurring [15].
Therefore, the most suitable encryption protocols for device
to server communications is AES.
1) Independent Key (Round Key)
The key will be arranged in the form of 4x4 bytes matrix
then it will be expanded into a schedule 44 words. The
independent key will be added into the input block of
plaintext to be ciphered once it is expanded [16].
2) Input Block
The first word from the key fills the first column of the
matrix. The expanded independent key is arranged into a
schedule of 44 words. Each round during the AES encryption
will contain 4 words from the schedule. Thus, the input block
contains a plaintext from the client side and expanded
independent key.
3) Input State Array
Before the round-based process for encryption commence,
the input state array is XOR with the first four words from the
schedule (44 words of expanded independent key).
C. Proposed AES Encryption Structure
1) White box cryptography (WBC)
Even though the implementation of white box will expose
an encryption algorithm to the external, white box is defended
by encoding, mixed bijections and external encoding. During
the encoding, a bijection of two keys are injected into the
lookup tables from the Sub-Bytes process. Apart from that,
mixed bijections will add confusion by concatenate input and
output from the encoding [16]. In addition, external encoding
will double the number of lookup tables. Although white box
provides full visibility of internal algorithm, it is just to
consider the worst-case attack model [3, 6].
2) Sub-Bytes/S-box
SubBytes dominates the AES performance. There are two
implementations to utilize the Sbox in SubBytes. The first
approach was implementing the composite-field
computation-based Sbox on FPGAs. Then another was the
LUT (LookUp Table). The former involves mapping and
inverse mapping between the GF (28) field and the GF (24)
field. While the latter one applies the formula, Lsb= 8 − K +
1, K is the number of inputs in LUT [4].
3) Shift Rows
This Shift Rows transformation is a simple permutation
(circular byte shift) process that will guarantee. For example,
in the first row, there is no byte to shift whereas at the second
row, there is a 1-byte circular left shift and followed by the
third row, a 2-byte circular left shift. Finally, the last row by
three bytes to the left. The shifts of row will help to
complicate the cryptanalysis that intrudes a cloud service.
The input block is in the form of column-wise. The first
four bytes of the input block will fill up the first column of
the state array, followed by the next four bytes of the second
column, and so on. In the end, the byte order of the input
block is scrambled up due to the rows shifted in the indicated
manner.
4) Mix Columns
In this step, combine the four bytes in each column of the
state with an invertible linear transformation. During the Mix
Column operation, four bytes are taken as input and then
outputs another four bytes, where each input byte influences
the four outputs. Along with the Shift Rows, another step in
the AES encryption, which is the Mix Columns, provide
diffusion in the cipher. A fast and low complexity
architecture for the Mix Column in AES operation is
proposed. It is recommended to contain a short critical path,
small gate count and versatility (encrypt and decrypt) [4].
5) Add Round Key
A round key will be derived from the original 128-bit
encryption key in every round. The XOR of the round key
(independent key) with the state array is a part of one of the
four steps available in both AES encryption and decryption.
The AES AddRoundKey algorithm is applied to derive the
128-bit independent key for every round from the original
128-bit encryption key. In addition, the logic behind the
independent key expansion algorithm is to guarantee that if
one bit of the encryption key is altered, the independent keys
for several rounds will be affected as well. Having said that,
the same manner goes to the 128-bit input block is arranged
in the form of state array, the algorithm organize the first 16
bytes of the encryption key in the form of 4x4 array of bytes.

Journal of Telecommunication, Electronic and Computer Engineering
58 e-ISSN: 2289-8131 Vol. 9 No. 3-11
6) Double the process of AES encryption in IoT with two
different keys.
If the hacker tend to send a malware that can interrupt,
inhibit the normal flow of data into and out of the system or
disrupt the IoT such as Smart Home System, the AES will
double encrypt by using two different keys, the malware will
unable to break the algorithm because it has been strengthen
by the double encryption process of AES [8]. Hence, the
system will be unable to disrupt by the hacker thus the user
able to access to the system as usual [6].
The data encryption and decryption above is for AES-128
bits. In order to make the data even more secure, we apply the
192 bits or 256 bits. Furthermore, in the near future, we can
produce the IoT board, which is a black box that can receive
the plain text as input and gives encoded output for various
IoT applications [4]. Having said that, even a common
computer user can secure data.
V. AES MATHEMATICAL ANALYSIS
A. White-Box Implementation
1) Encoding
The aim is to make the extraction of the key from the
encryption code harder due to a new complicated lookup
table. The bijection of key g and f.
T1 = g0 T0 f-1
2) Mixed bijection
A concatenation of input and output from the previous
encoding.
3) External Encoding
The double of lookup tables in the 1st to the 9th iteration
after the mixed bijection will produce:
228 of 8- bit to 32-bit lookup tables 1024 bytes each.
1728 of 8-bit to 4-bit lookup tables which is 128 bytes each.
16 of 8-bit to 8-bit lookup tables which is 256 bytes each.
Thus, the attacker will find it tedious to search through the
large storage for cryptanalysis.
Operations in AES are performed using bytes. There are
 number of bytes possible. Specific bytes are
represented by elements in the finite field,
  

where         (according to original
Rijndael algorithm). Other choices for p(x) can be used also.
B. S-box
S-box is used to transform a given element in   
 into
another unique element in F using the AES S-box.
For e.g.     
Binary presentation of   
Split it into two sides.
LHS
Binary presentation= 0010
Binary expression=   
Binary expression = 1
RHS
Binary presentation= 0110
Binary expression=   
Binary expression = 6
Entry 1 will be used as row and 6 will be used as column.
Referring to the AES S-box, we will gain 71, which as a byte
can be expressed as 010000111, and as polynomial can be
represented by.       
Thus, for the input     will produce output
   
C. SubByte
The input encoding of the i th round is offset by the output
of previous round and the encoding is composed of non-linear
mapping and linear mapping with the formula,
F = g. E. f
= Q. B. M. S. A. P
= QBMSAP
where: f = input encoding,
A = block diagonal linear mapping
P = nonlinear mapping
M = invertible linear mapping
S = the concatenation of S-box on m bits
E = 4F
Q = non-linear mapping
B = block diagonal linear mapping
Let x=(), S=(), AP=(),and M=[],
where ’s are 8-bit values, ’s are nonlinear bijections on 8
bits and  is the ith vertical strip of size 32 × 8. For 1 ≤  ≤
4,
F(x) = QBMSAP(x)
= (Σ=1 ())
= Q  (ΣQ−1 ())
The nonlinear part Q can be removed up to an affine
transformation in ((n/mQ)23mQ) when Q = (Q1,..,Qn/Qm)
and each Qi is a nonlinear bijection on mQ bits. After the
nonlinear P and Q are removed, the equation becomes F = B
S A. The specialized affine equivalence algorithm (SAEA)
with (n322n) is applied into the equation.  and  in O ((n/m)
mA323m) can be found, where mA is the smallest integer p
such that A is a block diagonal matrix with p×p matrix blocks.
Matrix block will be convert using S-box to produce a new
matrix.
 

D. ShiftRow
Let state be  where i=1, 2, 3,…, r, round=number of
rounds
 
 
After shift row,
 


Preventing DoS Attacks in IoT Using AES
e-ISSN: 2289-8131 Vol. 9 No. 3-11 59
E. MixColumn
This operating involve the modular multiplication of 2
four-term polynomials whose coefficients are element of
All the columns in state are treated as four terms
polynomial
       and
     .
Polynomial multiplication
   
 
   
where:
 
 
 
 
  
 
 
Modular reduction
Since the outcome of c(x) has 7-term polynomial, it has to
be reduced to 4 term, as it need to be in 4-byte word.
 
  
  
  
 
    
    
   
where:
  
  
  
 
Matrix representation
Using polynomial multiplication,
  
 
 
 
Replace coefficients of a(x) with the constants [3 1 1 2], we
will get:
  
   
  
   Or,
 

F. AddRoundKey
Each Round Key consist of Nb words from the key
schedule are each added into the column of the state, such that
[] [] • [] for 0 • c
< Nb.
The key matrix  where i=1,2,…,r-1 will be added to
matrix  resulted from MixColumn. This will produce 
where i=1,2,…,r-1.
 
For the first round i=1,
For i=r, formulae  
where  was output of ShiftRow,
=cipher text
is used because MixColumn transformation is not repeated in
final round.  . This proves that our proposed method is
more secure in terms of avoiding the attacks especially the
DoS attack. It can be used in IoT setting anywhere, as it will
ensure that Man-in-the-Middle attack will not occur. In
future, we plan to test our algorithm on IoT devices and check
its security and execution time to benchmark it with other
security algorithm.
VI. CONCLUSION
This paper presented novel AES implementation with a
simple and integrated countermeasure against Denial of
Services Attack (DoS) in IoT. IoT can be divided into smart
grid, smart car, smart campus, smart house, wearables and
industrial internet. The security involves the measures and
controls that ensure confidentiality, integrity and availability
of the information processed. However, there are some of the
attack techniques of DoS such as Attack Tools, Application-
layer Floods, Degradation-Of-Services and Denial-Of-
Services. Rijndael algorithm incorporates four
transformations namely SubBytes, ShiftRows, MixColumns
and lastly is AddRoundKey in each round step and it is a
uniform and parallel composition. AES has a strong
symmetric key cryptographic algorithm, which uses table of
lookups. To carry out the encryption and decryption, AES
need a minimum of 128 bits or 192 bits with recommendation
of 256 bits cryptographic keys. The algorithms is composed
of three layers, which are linear diffusion, non-linear
diffusion and key mixing. The new design and idea allows the
construction of the actual cores with speed characteristics and
efficient area. Protecting against the threats and attacks in the
cyber world still need to be investigate and excavation
however it is still a challenge and it is also costly.
Development of the efficient countermeasures for the
prevention of the side-channel attacks in IoT is one of the
open issue in scientific world and it should be further
investigated optimized in the future research by human. The
proposed algorithm will be compared against present security
algorithms for security effectiveness and their effect on
memory as well as computational usage.
ACKNOWLEDGEMENT
The work is fully funded by Research and Innovation
Management Center (RIMC) under the grant number
F08/SpSG/1403/16/4.

Journal of Telecommunication, Electronic and Computer Engineering
60 e-ISSN: 2289-8131 Vol. 9 No. 3-11
REFERENCES
[1] J. Gubbi, R. Buyya, S. Marusic, and M. Palaniswami, “Internet of
Things (IoT): A vision, architectural elements, and future
directions,” Future generation computer systems, vol. 29, no. 7, pp.
1645-1660, 2013.
[2] B. Li, J. Yub, “Research and application on the smart home based on
component technologies and Internet of Things,” Advanced in Control
Engineering and Information Science, vol. 15, pp. 2087 – 2092, 2011.
Retrieved from http://www.sciencedirect.com.
[3] M. A. Zaveri, S. K. Pandey, and J. Kumar, “Collaborative service
oriented smart grid using the Internet of Things,” In Communication
and Signal Processing (ICCSP), 2016 International Conference on, pp.
1716-1722, Apr. 2016. IEEE.
[4] S. Pawel, “FPGA Trojans through detecting and weakening of
cryptographic primitives,” IEEE Transactions on Computer-Aided
Design of Integrated Circuits and Systems, vol. 34, no. 8, pp. 1236-
1249, 2015.
[5] A. A. Panmu, K. S. Chong, W. G. Ho, and B. H. Gwee, “Interceptive
Side Channel Attack on AES-128 Wireless Communications for IoT
Applications,” Asia Pacific Conference on Circuits and System 2016,
pp. 650-653, 2016.
[6] G. Singh, “A study of encryption algorithms (RSA, DES, 3DES and
AES) for information security,” International Journal of Computer
Applications, vol. 67, no. 19, 2013.
[7] M. M. Wong, M. L. D. Wong, A. K. Nandi, et al., “Composite field
GF(((22)2)2) Advanced Encryption Standard (AES) S-box with
algebraic normal form representation in the subfield inversion,” IET
Circuits Devices Syst., vol. 5, no. 6, pp. 471–476, May 2011.
[8] L. Whitney, Adobe hack attack affected 38 million accounts, 2013.
Retrieved from CNET: https://www.cnet.com/news/adobe-hack-
attack-affected-38-million-accounts/.
[9] C. Paar, J. Pelzl, and B. Preneel, Understanding Cryptography: A
Textbook for Students and Practitioners, 2010. Springer.
[10] A. Jose, “SecCloud Protocol Implementation Using AES Algorithm
for Security and Privacy in Cloud Computing,” 2014 International
Journal for Research in Applied Science and Engineering Technology,
vol. 2, no. 2, pp. 1-8, 2014.
[11] J. Yan, F. Chen, “An Improved AES Key Expansion Algorithm,”
International Conference on Electrical, Mechanical and Industrial
Engineering, 2016.
[12] A. Al-Haj, G. Abandah, and N. Hussein, “Crypto-based algorithms for
secured medical image transmission,” IET Information Security, vol. 9,
no. 6, pp. 365-373, 2015.
[13] S. Srinivas, “FPGA Based Hardware Implementation of AES Rijndael
Algorithm for Encryion and Decryption,” International Conference on
Electrical, Electronics, and Optimization Techniques, pp 1769-1776,
2016.
[14] N. Ruan and Y. Hori, “DoS atack-tolerant TESLA-based broadcast
authentication protocol in Internet of Things,” International
Conference on Selected Topics in Mobile and Wireless Networking, pp
60-65, 2012.
[15] Y. Yuan, H. Yuan, L. Guo, H. Yang, and S. Sun, “Resilient Control of
Networked Control System Under DoS Attacks: A Unified Game
Approach,” IEEE Transactions On Industrial Informatics, vol. 12, no.
5, pp. 1786-1794, 2016.
[16] N. Mathu and R. Bansode, “AES Based Text Encryption Using 12
Rounds With Dynamic Key Selection,” 7th International Conference
on Communication, Computing and Virtualization 2016, pp.1036-
1043, 2016.
[17] Q. Liu, Z. Xu, and Y. Yuan, “High throughput and secure advanced
encryption standard on field programmable gate array with fine
pipelining and enhanced key expansion,” IET Computers & Digital
Techniques, vol. 9, no. 3, pp. 175-184, 2015.