Hybrid Red Deer Algorithm with Physical Unclonable Function for Security Enhancement in IoT-WSN

International Journal of Computer Networks & Communications (IJCNC) Vol.14, No.5, September 2022
Hybrid Red Deer Algorithm with Physical
Unclonable Function for Security Enhancement in
IoT-WSN
Shilpa Venkata Rao and Vidya Anantha
Department of Computer Science and Engineering, Vivekananda Institute of
Technology, Bangalore, Karnataka, India
Abstract. In this modern era, Wireless Sensor Network (WSNs) have turned out to be a more attractive
option, because of the advancements in communication. If the network is insecure, an attacker can intercept
messages and break the sensor nodes’ security; as well as duplicate the authentication codes to launch a
variety of attacks. As a result, Physical Unclonable Functions (PUFs) with unpredictable features are
encouraged to be used in the development of lightweight cryptographic protocols. This work introduced
the Red Deer Algorithm (RDA) with PUF as a new mutual authentication system. When a sensor receives
a challenge from the gateway, the inbuilt PUF generates a key and distributes it to the sensor by providing
complete resilience against malicious attacks. PUF is resistant to node acquisition, cloning, and malicious
attacks, as well as node physical security flaws. Key distribution is moving too quickly and the adversary
won’t be able to conduct a harmful attack in time. Furthermore, PUF’s unclonability and unexpected
qualities provide key uniqueness and two-way authentication for improving the security. When compared
to existing Tunicate Swarm Grey Wolf optimization (TSGWO) and PUF-Based Mutual-Authenticated
Key Distribution (PUF-MAKD), the proposed RDA-PUF demonstrated better results. The simulation
results are obtained in terms of minimizing energy consumption as 0.7 J, end-to-end delay as 7 sec, packet
delivery ratio of 97%, increasing network lifetime to 1330 sec, and improving secure connectivity to 1.211.
Keywords: Network Lifetime, Physical Unclonable Function, Red Deer Algorithm, Security, Wireless
Sensor Network.
1 Introduction
Wireless Sensor Network (WSN) is now a social transmission network with sensor nodes,
mortal structures and Base Stations (BSs) [1]. Internet of Things (IoT) develops inventive
ideas and advances a variety of communication formats. IoT structures rely on WSNs
to collect data, and cloud/edge/fog computing can boost their computational dimensions
[2], [3], [4]. Several IoT services and applications have been provided in this manner. IoT
offers to make the Internet more pervasive, with the potential to influence several aspects
of consumers’ entertainment [5]. Because of insufficient sensor techniques, exploitation of
energy throughout the routing procedure may cause a problem. Several hops are directed
at the vast possibilities of land section over a few wireless machines [6]. WSN is a critical
component of the Internet of Things, and it has evolved into a variety of different practical
applications [7]. WSNs and the Internet of Things (IoT) now have a variety of precarious
and non-critical manifestations that affect nearly every aspect of an individual’s daily life
[8]. WSN nodes are frequently tiny and rely on battery-powered technologies [9]. As a
result, energy-efficient ways extend the network’s lifespan, which is critical. There are sev-
eral approaches and procedures for implementing an energy-efficient notion in WSN-based
IOT architectures [10], [11].
Many exploratory projects have focused on the design of routing/ clustering procedures
and additional schemes [12] for energy saving in recent years, but only a few have accom-
plished to model the systems that deliberate the construction of clustering and routing
DOI: 10.5121/ijcnc.2022.14503 31

mutually [13], [14]. WSNs’ Quality of Service (QoS) must be improved in the current cir-
cumstances because the system must last for a long time [15]. A structure using effective
clustering/routing via optimization approach [16] is another source of motivation, because
the lack of network resources, necessitates the task of creating an energy-efficient trans-
mission path and set-up design [17]. Researchers have developed different authentication
techniques based on a PUF in a Radio Frequency Identification (RFID) system to avoid
clone attacks. However, most protocols require verifiers to connect to a database con-
taining a high number of PUF Challenge-Response Pairs (CRPs), which is not feasible for
memory-constrained WSN nodes [18]. The current study focuses on the clustering/routing
method developed for WSN-based IoT presentations. The routing method is shown as an
important aspect of transferring the data gathered from sensors to the end point for fur-
ther processing. After the data pre-processing is done, it is possibly used in a variety of
applications to help people develop in their lives [19], [20].
The paper represents RDA-PUF based clustering and routing for improving the network
lifetime. The major contribution of this research are
– Initially, an RDA-PUF based clustering method is created using an effective particle
encoding design and its fitness derivative.
– Then, combining the RDA-PUF with a random deployment model for large-scale sensor
networks. A sensor is assigned to gateways to promote secure connectivity and protects
a sensor using PUF.
– RDA-PUF based routing procedure is processed among the transmission distance and
amount of data transferred for comprehensive multi-objective function.
– RDA-PUF is simulated and compared to other important distribution schemes in terms
of energy consumption and security.
In existing methods, the major issue was presented in the design which are incapable
to analyse the validity of received message and every CRP is exploited just once to avoid
the attacks. In order to overcome these problems, proposed RDA-PUF is designed with
resource-constrained devices which validates the message through PUF technology and
derive a shared session key for secure communication. The structure of this research are
mentioned as follows,
Section 2 describes the exiting studies based on clustering/routing and security in WSN-
IOT. Section 3 defines the problem statement along with solution. Section 4 explains the
preliminaries of PUF, energy model and overview of RDA. Section 5 defined the process of
clustering/routing using proposed RDA-PUF, working process of key authentication and
PUF security in detail. Section 6 demonstrates the simulations results and its comparative
analysis. Finally, conclusion is declared in section 7.
2 Related Work
Wireless Sensor Network-Internet of Things (WSN-IOT) makes use of a variety of current
procedures, methodologies, and ideas from traditional wireless networks. However, some
significant changes necessitate the use of novel techniques and methods. In this segment,
a few of the gathered studies on clustering and routing in WSN-IOT are briefly discussed.
PUF-Based Mutual-Authenticated Key Distribution (PUF-MAKD) for Dynamic Sensor
Networks has been proven by YananLiu et al. [21] to assist the sink node in authenti-
cating and distributing session keys to static and mobile sensors. A challenge-response
32

technique based on the PUF ensures lightweight mutual authentication. To combat the
PUF Challenge-Response Pairs (CRPs) exposure problem, the CRPs are not sent in plain-
text to protect PUF from a modelling attack. Furthermore, sensors are not required to
recover any password, reduces loading overhead and increases sensor node attack resilience.
However, the PUF response was not sent in plain form to avoid modelling attacks, which
was exploited in other existing PUF-based techniques.
For WSN-IOT hot-spot problems, Mohammad Ali Alharbi et al. [22] developed an energy-
efficient clustering/routing resolution. This clustering technique uses a basic Cluster Head
(CH) selection method to choose a CH based on aggregate node communication loads and
remaining batteries, and it is dependent on the placement of WSN nodes. As a conse-
quence, offer a routing technique that assigns marginal connections to the selected cluster
leader to address weak situations. On the other hand, the suggested process only achieves
the highest performance requirements for 2–3 communication nodes only, after that it
slows down automatically.
For WSN-IoT, IpekAbastkeles-TurgutGokhanAltan [23] presented the Fully Distributed
Energy-Aware Multi-level (FDEAM) approach. And also, an inter-cluster transmission
process is introduced to enhance efficiency. Furthermore, the coverage region of the second-
level gathering was actively developed, offering BS cluster distance. The FDEAM technique
was used by cluster heads without a primary appliance to choose clustering boundaries.
FDEAM technique was a reasonable starting point, even though it is useless for non-
uniform node assignments and is reliant on a particular source.
In WSN-IoT, KavitaJaiswal and VeenaAnand [24] designed an Energy-Efficient Optimal
Multipath Routing (EOMR) algorithm to improve QoS metrics. The recommended EOMR
display generates an imbalance in the network due to greater bandwidth. When finding
the ideal path, the suggested method takes into account some characteristics, including
dependability, lifespan, and increased traffic during the next node. Here, transmissions
occurred regularly in some areas of the system, while others were used in all the regions.
The Tunicate Swarm Grey Wolf optimization (TSGWO) approach was used by Nitesh
Chouhan and Jain [25] to create the multipath procedure. TSGWO generates numerous
routes from a distinct node to several endpoint nodes. The source node immediately trans-
fers data packets to the target point. The optimal path was the one with the lowest number
of delays, the lowest routing distance, and the highest connection durability. Route con-
servation was also allowed for route splintering retrieval through DRINA in the event of
a link failure. But this method ignores the servicing parameters.
Shilpa V and Vidya A [26] has provided secure transmission of data and mutual authen-
tication for IoT-WSN using Hybrid Optimization Algorithm (HOA) and Shamir Secret
Sharing (SSS) method. The arrangement of Grey wolf optimization (GWO) and Moth
Flame Optimization (MFO) was termed as HOA where the fitness functions were calcu-
lated. Moreover, SSS was utilized for delivering mutual authentication amongst the nodes.
The Cluster Head (CH) selection and optimal routing eliminates the malicious nodes that
advanced to reduce the packet loss in the network.
Shilpa V, Vidya A, Santosh Pattar [27] has presented dependable secure transmission of
data among IoT were attained by lightweight encryption method. Here, a Secure Reliable
Message Communication (SEC-RMC) procedure by means of Mosquitto MQ Telemetry
Transport (MQTT) message broker through cryptographic developments to provide secu-
rity against hacking and packets exploitations. The suggested SEC-RMC method mini-
mizes the amount of messages transferred amongst the Internet of Things (IoT) devices.
33

Li et al [28] has demonstrated a provably secure and real-world PUF dependent end-to-
end mutual authentication and key exchange technique for IoT. This research suggests a
scheme for IoT through integrating PUF with Certificate Less Public Key Cryptography
(CL-PKC) on elliptic curve to improve the security. This process merely demands multiple
handshakes without including the practical server contribution to deliver accurate secrecy.
However, this method does not create the shared session key among applicants and public
key encryption process to confirm the secrecy.
3 Problem Statement
– The key problem in routing process was losing connectivity during data transmission
while utilizing energy control mechanisms.
– PUFs was assumed as one-way hardware functions that are simple to implement but
difficult to clone, reproduce, or predict.
– Data loss has occurred due to a faulty network, therefore, data dependability must be
monitored during data packet transfer.
– The involvement of various objects and nodes energy become important concerns in
WSN, because of insufficient energy given to the nodes.
Solution: This research’s main purpose is to provide energy-efficient clustering and
routing for data transmission via the Wireless Sensor Network-Internet of Things (WSN-
IOT). PUF can withstand node acquisition, cloning, and malicious attacks, as well as
entirely prevent node physical security vulnerabilities. And key distribution development
is too fast, so the attacker won’t be able to launch a malicious attack in time. The proposed
RDA-PUF architectural framework prevents the system from data loss, unbalanced load,
inflexibility, and unsecured connectivity.
4 Preliminaries
The major goal of this study is to provide energy-efficient clustering/routing for data
transmission over the Wireless Sensor Network-Internet of Things (WSN-IOT). The sen-
sor would not have any keys pre-loaded in its storage, however, when it receives a request, it
generates a PUF response on the fly to decrypt the key supplied by gateway. Load balanc-
ing, flexibility, and secure connectivity are all possible with the conceptual methodology.
A current transmitting process must be accessible and adaptable to complete network
changes.
4.1 Physical Unclonable Function (PUF)
PUFs are innovative circuit elements that extract information from the physical properties
of Integrated Circuits (ICs). There are two parts to PUFs: a physical component and an
operating component. The physical aspect is a difficult-to-replicate sophisticated physical
system. Its unclonability is due to unpredictably fluctuating manufacturing techniques [29].
Within the instance of PUFs resting on IC, manufacturing modifications are frequently
influenced by deep-submicron fluctuations. To convert the dynamic system into a service
that can be used for identification, a set of relevant inputs and a set of suitably variable
answers has to be provided.
Figure 1 shows the general working function of PUF. From the figure 1, there are four
major components are presented, they are, Cluster Head (CH), Base Station, Sensor node
34

Fig. 1. General working process of PUF
and PUF-key distribution. To prevent clone attacks, researchers have explored various
security approaches based on a PUF concept. Many algorithms, unfortunately, necessitate
verifiers connecting to a library that contains a collection of PUF Challenge-Response
Pairs (CRPs), which is not possible for memory-constrained Network elements. It clearly
observed that the information from source to destination are passed through the Cluster
Head (CH). If in case, PUF key distribution acknowledge the CH1 as leader, sensor node
passes through CH1 to destination.
4.2 Energy Model
The energy model and its usage in Sensor setup has been defined in this subsection. The
architectural system of WSN is mostly based on dualistic communications infrastructure
characteristics, such as a multipath channel [30]. The bandwidth decreases with time as
the sample free space resource is consumed, necessitating the use of a multi-path com-
munication. The energy model’s structure for source/destination is mathematically stated
and presented in the equation (1),
EnerT
(1, D) =

L∗EElec + 1∗ ∈FS ∗D2, D D0
L∗EElec + 1∗ ∈MP ∗D2 D ≥ D0

(1)
The prerequisite energy in transmission networks is signified as ∈MP , if it is in free
space, it is signified as ∈FS , whereas, in position of receiving channel it is nominated as
EElec . The whole energy is essential for data communication during the distance d which
is premeditated after eqn. (2). The receiver’s demanded energy is formulated as eqn. (2),
EnerR
(L) = L∗EElec
(2)
Receiver’s required energy is characterized as Ener(L) distance amongst the sender
and receiver is signified as ∈MP ∗D2. ∈FS ∗D2. EElec is designed in the course of several
characteristics.
4.3 Overview of Red Deer Optimization
Red Deer Algorithm (RDA) starts with a random population of search agents that is spread
evenly throughout the population [31]. To assign the male red deer, the sensor nodes with
the best potential in terms of residual energy, cluster head distance, and cluster to base
station distance are picked, with the remainder nodes referred to as hinds. The sensor
nodes having the best chance of being chosen as CH are male red deer. The hinds are
International Journal of Computer Networks Communications (IJCNC) Vol.14, No.5, September 2022
35

cluster members who serve as sensor nodes and are completely under the direction of the
male red deer nodes.
Initial Population Generation The parameters considered for the most important
optimization goals in establishing the global solution or near-optimal solution, are defined
in equation (3). When the CH selection procedure is seen as a d-dimensional optimization
method, the red deer indicates the number of d-dimensional arrays that range from 1 to
d. Equation (3) is used to define red deer array as
RD(Array) =

FS(1), FS(2), FS(3) . . . ..FS(d)

(3)
Where, RD(Array) signifies the individual dimensions. Subsequently, the fitness function
values are evaluated for every RD which is expressed in Equation (4)
FitvalueRD(Array) = FitFn

FS(1), FS(2), FS(3) . . . FS(d)

(4)
Primarily, Psize is stated as the size of the initial population which establishes the
execution of the method. The best male RDs (RDmale) and the residual exploration solu-
tions are elected as RDHinds = Psize − RDmale. Additionally, the amount of RDmale and
RDHinds supports in preserving the amplification and diversification.
Roaring Stage In particular, the male RD solutions constantly modify their location
which is updated as shown in Equation (5).
RDmale =

RDmale−old + UDUDR(2) × (UTH − LTH) ∗ UDR(2) + LTH

if UDR(3) ≥ 0.5
RDmale−old + UDR(1) × (UTH − LTH) ∗ UDR(2) + LTH

if UDR(3) .05

(5)
The current position of male RD is stated by RDmale−old; UTH and LTH depict the up-
per and lower search space respectively. The random variables are UDR(1), UDR(2), UDR(3)
which are produced by means of uniform distribution (0 and 1).
Male Leader assortment from male red deer The male RD solutions are divided
into two categories at this phase: leader and RD solutions. The proportion of male leader
responses is estimated at this point using Equation (6).
RDComd−Count = Round (β, RDmale) (6)
RDComd−Count signifies the amount of leader male in the clusters. β is the initial value
(0, 1) connected with the percentage of solutions measured for utilization. At this stage,
the stag RD solutions (RDstag−Count) are considered in Equation (7)
RDstag−Count = RDmale − RDComd−Count (7)
Stag and Male commanders fighting phase Here, one superior solution is determined
in comparison to the leader and two additional solutions are discovered by following the
procedure. Equation (8) and Equation (9) are used to illustrate the fighting process quan-
titatively.
NewSol(1) =
SolComd + SolStag
2
+ UDR(1) × (UTH − LTH) ∗ UDR(2) + LTH

(8)
36

NewSol(2) =
SolComd + SolStag
2
+ UDR(1) × (UTH − LTH) ∗ UDR(2) + LTH

(9)
Where, UTH and LTH represents the upper and lower search space. Moreover, UDR(1)
and UDR(2) are also random variables generated using a uniform distribution. In this
fighting stage, the first solution NewSol(1) selects the success solution and the second
solution NewSol(2) helps in identifying losing solution.
Construction of Clusters During the cluster construction stage, the male leader answer
becomes the cluster head, which controls the group of hinds. The amount of hind solutions
are proportionally allocated among the commander solutions to build clusters as shown
in equation (10).
Nval = Cmdpower(n) − Max Cmdpower(i)

(10)
Where, Nval and Cmdpower(n) signify the normalized mean of leader and control of
every entity leader, respectively, at every round of execution. After that, the leaders’
normalized power is designed which is shown in Equation (11).
RDComd(Power) =
CmdPower(n)
PRDComd−Count
i=1 CmdPower(i)
(11)
Where, RDComd(Power) signifies the power infatuated by every leader for the duration of
clustering and CH selection. For the duration of the clustering process, the number of hind
alternatives can be controlled by the commander’s solutions. Equation (12) is exploited to
calculate the number of hinds that might conceivably occur in a cluster (harem).
RDHind−Harem = Round (α, RDHind−Cnt) (12)
Where, RDHind−Harem represents the number of hinds present in a cluster, for the
duration of CH selection. Overall, the parameter with a higher fitness function is referred
as leader which is found to be superior to the hind options. In the operation of clustering
in WSNs, the commander and hinds symbolize the CH and cluster members.
5 Proposed Method
RDA-PUF is meant to reduce the incremental energy usage at every node in this study.
Initially, each node is created via similar energy sources. Energy is reduced during the
period of communication and dictated by the development of each node. The overall
structure of the suggested RDA-PUF is explained in Figure 2.
Fig. 2. Overall Structure of Proposed RDA-PUF
37

Step 1: At first, the network parameters are initialized with sensor node deployments
in the network.
Step 2: Then the nodes are randomly positioned in the related region, after that mobile
nodes are designated as energetic which is exclusively dependent on node location.
Step 3: The calculation of fitness function parameters are estimated.
Step 4: RDA-PUF is used for secure clustering, CH selection and routing process to
find the shortest path.
Step 5: Routing process is created by means of proposed RDA-PUF which are exploited
to develop the optimal route among CH and BS.
Step 6: Whenever the route from source to destination is recognised, the source node
transmits the data in destination.
Step 7: After the data transmission, RDA-PUF based mutual authentication gets pro-
cessed.
Step 8: Then RDA-PUF estimates the optimal path by taking various objective func-
tions.
Step 9: BS is regularly exploited to detect the nodes energy. In order to eliminate
secure in connectivity, energy consumption and data loss re-clustering/rerouting is done
repeatedly.
The functionalities of each step and its requirements including CH selection, Fitness
function metrics, route selection, PUF structure, Key distribution, Authentication and
security along with flowchart for proposed RDA-PUF is explained clearly in the below
sections
5.1 CH selection
The RDA-PUF is used for CH selection and routing. The suggested RDA-PUF is chosen
since it has a collision prevention characteristic by default, making it particularly useful
for generating congestion-aware routing in unexpected scenarios.
Fitness function The trust metric, Queue Utilization between Nodes (QUN), link qual-
ity, and distance are all employed in the RDA-PUF technique. As a result, the specified
cost measurements are employed to prevent malicious nodes, by giving a major focus to
the trust value for the trust metric, throughout the routing. To increase the presentation
in terms of energy consumption and overhead, subcategories of the cost indicators like
QUN, link quality, and distance are included.
Trust metric: Trust is a primary cost value for routing via an IoT-based WSN, and has
been used to increase security against malicious nodes. The mobile nodes in an IoT-based
WSN interact with one another by relying on the mutual trust established between the
nodes over time. Direct trust is a packet-forwarding behavior that is based on the fraction
of data (transferred) (TPij) and the amount of data packets (received) (RPij) . The direct
trust degree is calculated based on the following equation (13).
Trust =
TPij
RPij
(13)
QUN and link quality: QUN is demarcated as the ratio amongst the amount of packets
engaged in the queue of jth node (QUNj) and quantity of obtainable packets in the
particular node’s queue (QUNtotal). Equation (14) is utilized to calculate the QUN.
38

QUN =
QUNj
QUNtotal
(14)
Link quality is described as the proportion amongst the quantity of transmission and
retransmissions that are necessary for successful data transmission among node i and j .
Linkquality =
1
f × r
(15)
Where f and r state the forward reverse information delivery respectively.
Distance: Euclidean distance among the nodes are premeditated to recognize the direct
route over the IoT-WSN and deliberated through the equation (16).
Distance =
q
(xj − xi)2
(yj − yi)2
(16)
Where, xi, xj and yi, yj are coordinates of nodes i and j. In RDA-PUF, a weight value
(δ) is considered for each cost value to convert into a single cost value as shown in the
equation (17).
ck = δ1 × Trust + δ2 × QUN + δ3 × Link quality + δ4 × Dis tan ce (17)
0.35, 0.25, 0.25 and 0.15 are the weighted mean for δ1, δ2, δ3, and δ4 correspondingly.
The malicious nodes that generate packet drops are avoided using the trust metric when
designing the transmission path. Following that, to generate a route with less traffic,
the QUN and link quality are taken into account in the cost measure. This reduces the
chance of a communication collision and improves packet delivery. In addition, the shortest
distance is determined to reduce the nodes’ energy usage.
Route Selection When a client gets data in an IoT-based WSN, the client registers its
ID in the routing table. The client begins forwarding the data to the desired destination
after receiving it. The RDA-PUF evaluates the route with the highest pheromone rate if
the source consumer receives multiple detection notifications from its neighbors. Direct
transmission of data is performed by clients having a one-hop connection to the router,
while multichip communication is performed by clients further away from the router via
the RDA-PUF. As a result, data packets are transmitted along the path with the high-
est signal level. And once the routing path is established utilizing RDA-PUF, the data
packet distribution is started. The queue length employed in routing is utilized to create
a collision-free path that results to data being transmitted over a network with reduced
congestion. The link quality is then utilized to determine the route’s mean based on the
sent and received packets. As a result, above-mentioned fitness parameters are exploited
to determine the best data transmission route.
5.2 PUF Structure
Rather than keeping secrets in storage, PUF-based components are used to produce secrets
from randomized variability in the production of electronic components. They don’t ne-
cessitate costly cryptographic hardware like the Secure Hash Technique (SHA) standard.
Because the ”secret” is obtained from substantial features of the Integrated Circuit (IC),
even with full comprehension of the semiconductor manufacturing, two equal transistors
cannot be manufactured. PUF components would be investigated as a replacement that
effectively uses a device’s different characteristics as a hardware-based signature owing to
39

manufacturing variances. There are electro - optic PUFs and coating PUFs in addition to
IC-based PUFs.
Every PUF has been treated as black-box ranking scheme, in which the system is presented
with a number of challenges and returns with a collection of suitably varied solutions. PUF
is described by a one-way transformation matrix, which may be written as Eq. (18).
Γ : R → Y : Γ(x) = y, x ∈ X, y ∈ Y (18)
In a PUF function, X and Y stand for challenge and response sets, respectively, in Eq.
(18). For the given PUFΓ, a Challenge-Response Pair (CRP) (x, y) is formed by a detailed
trial x and a matching response y. Because of qualities such as low-weighted, unclonable,
and unpredictable, PUFs are used for two main purposes: low-cost validation and secure
key creation. Subsequently, a sensor-based wearable PUF was developed, which makes use
of different strengths supplied by a variety of sensors found on such devices. Low-level
accessibility can be implemented during the production process or by the installation of a
particular device controller to exploit such distinctive featured equipment.
RSA Cryptography RSA public key cryptosystem was established by Rivest, Shamir
and Adleman, that is extensively exploited for secure data transmission. This research
assesses the development of a number of symmetric and asymmetric encryption methods
known as RSA in order to determine whether the encryption ratio is high when both
encryption techniques are used. As per various criteria, every method has its own benefit.
Because the key length in an asymmetric encryption technique is long, breaking the code in
RSA is difficult. When it comes to throughput, more throughput means less power usage.
RSA algorithm is stated as the better option in symmetric key encryption approaches
and more secure in the asymmetric encryption technique because it generates keys by
factoring high prime numbers. As a result, RSA is determined to be the superior answer
in this manner. The comprehensive description of RSA is specified as follows.
Key Generation
Step 1: Initially, select two dissimilar and large primes pq and calculates n = p × q
and Euler’s totient function is represented as φ(n) = (p − 1) × (q − 1)
Step 2: Select a public key parameter e (i.e.) 1 e φ(n) and gcd(φ(n), e) = 1.
Step 3: Then, calculates the private key parameter d (i.e.) (e×d)−1 = 0 mod φ(n) by
means of Extended Euclidean procedure. (e, n ∈ Z × Z∗
φ(n)

and

d, n ∈ Z × Z∗
φ(n)

are
stated as public/ private key pairs.
Encryption The function of encryption eh : Zn → Zn is stated as
eh(M) : b = Me
mod n
Decryption The function of decryption dh : Zn → Zn is stated as
dh(M) : M = bd
mod n
Steps for Key Distribution
– Initialization
Ahead of deploying the network, every sensor S is drive-in PUF, as stated by ΓS. C
is referred to as an arbitrary challenge number that acts as input for ΓS and receives
the response R (output). The cycle is stored as id, C, R into G (gateway).
40

– Distribution of Key
Once deployed in the network, G states the C from the list and directs it to S. It
proceeds C as input and becomes the equivalent R as Eq. (19).
R = ΓS(C) (19)
S utilizes R to convert C then drives to G:
Cipher = E(R, C) (20)
From equation (20), encryption key is referred as R, plaintext is referred as C, and
cipher is referred as cipher text. G utilizes R to decode and associate through C:
Plain = D(R, cipher) (21)
From the eqn. (21), the decryption key is referred as R which is obtained through S.
If Plain = C is true, G trusts S and is validated for the duration of Initialization.
At that time it produces key kGS and directs S through an encryption process of R,
which is mentioned as Eq. (22):
Cipher2 = E (R, kGS) (22)
S decrypts Cipher2 and obtains kGS.
Or else, if Plain = C remains incorrect, G deliberates through an unauthenticated
device which will make it incapable to create a session key by means of a sensor.
Authentication The CRP id, C, R of every approved device with a PUF is saved
by the gateway. When the sensor receives a challenge C, the PUF creates a response R
and directs the Ciphertext Cipher = E(R, C), and compares it against challenge Plain =
D(R, cipher), because only authorised sensors can deliver the right answer R in response to
C. Then it transmits an encrypted assembly kGS with R. Since only authorized gateways
are familiar with the sensor’s accurate PUF CRP sequence, this procedure aids the sensor
in authenticating the gateway. The attackers might provide collision attacks on the entire
network by accessing the passwords or other confidential documents in proportion to the
captured nodes’ memory. As the key distribution development is too fast, the opponent
won’t be able to launch a malicious attack on time.
PUF Security PUF’s security is based on physical peculiarities that are impossible to
replicate even at the industrial-technological sophistication. PUF must be incorporated
into a node’s chip. To put it another way, even if the attacker is aware of the PUF ar-
chitecture and some difficulties, they will be unable to generate the appropriate replies.
PUF can withstand node acquisition, cloning, and malicious attacks, as well as entirely
prevent vulnerabilities of node physical security. Figure 3 shows the flowchart of proposed
RDA-PUF method.
The flowchart steps are specified as follows:
Step 1: Initially, the sensor node is established.
Step 2: Once the node gets initialized, the process of roar male RDs is started.
Step 3: After that, select the energy efficient nodes (male commanders).
Step 4: Then harem formation is created.
Step 5: Mate is staged with the nearest hind to find the shortest path.
Step 6: Then the mobile node is selected for next set of generation.
41

Fig. 3. Flowchart of Proposed RDA-PUF
Step 7: Identify the transmission path
Step 8: Check the stopping criteria
Step 9: IF No, go back to step 2; otherwise, select this optimal path for further trans-
mission.
Step 10: Check the selected path with key distribution process.
Step 11: Following the clustering/routing procedure, the PUF-based mutual authenti-
cation mechanism can be examined to prevent malicious attacks.
Step 12: Examine the PUF’s delivered performance parameters
Step 13: Stop the process.
Data Integrity Verification The trust value computation is used to validate and verify
data integrity in WSN in this study. The observed information is secure utilizing shared
symmetric encryption in this step. The encrypted data is then broken up into smaller
chunks. After that, every one of these blocks is signed with a homomorphic tag and sent
to the parent block, which has a higher trust value. The signed blocks are forwarded to
the aggregator once they have been obtained. The sensor additionally gives the aggregator
information about the parent block, as well as the block numbers. After confirming the
accuracy of the data and confirming the integrity of every block, the aggregator applies
the aggregation function to them. The combined output is then sent to the sink, in which
it is double-checked by the sink. For the sensed data, this procedure is iterated by every
sensor. Those findings were supported using the detection ratio in relation to node counts,
as detailed in section 6.7 below.
6 Results and Discussions
The outcomes of developed RDA-PUF for multicast routing in an IoT-enabled WSN are
discussed in this section. The conclusions and recommendations of the suggested RDA-
PUF technique are fully described in detail. The 4-GB RAM and Intel Core CPU are used
42

by the RDA-PUF. The proven RDA approach is utilized to provide secure transmission
between the source and the BS. Table 1 displays the design and simulation of the RDA-
PUF technique under different parameter variables.
Table 1. Requirement of model parameters
Parameters Value
Number of Nodes 400
Packet Size (bytes) 512
Receiving power (W) 1.0
Simulation Time (s) 100
Traffic type CBR
Transmission power (W) 1.4
Data level (Kb/s) 300
Initial energy (J) 100
MAC layer IEEE 802.15.4
Network Size (m) 500 × 500
6.1 Performance of Energy Consumption
The study of projected RDA-PUF and associated systems, such as FDEAM [23], EOMR
[24], and TSGWO [25], in the context of energy consumption, is shown in figure 4. Figure 4
shows that when compared to the remaining three conventional approaches, the suggested
RDA-PUF consumes lesser energy of 0.7 J. A node with less value is twisted by the packet
boosting stage in the RDA-PUF approach. Furthermore, more energy is preserved by
means of highest optimality factor continues to be important in progressing data packets;
while, the traditional methods require an extra node when dispatching the same data
packet; as a result, sophisticated energy is exhausted in conventional techniques.
Table 2. Performance of Energy Consumption
Number of Nodes Performance of Energy Consumption (J)
FDEAM [23] EOMR [24] TSGWO [25] Proposed RDA-PUF
100 0.8 1 8 0.5
150 7 9 15 6
200 10 14 19 8
250 14 19 26 13
300 26 30 31 20
350 29 36 37 23
400 41 47 44 29
The comparative examination of energy consumption performance is tabulated in Table
2. According to table 2, the developed RDA-PUF consumes less energy when compared to
existing approaches. Figure 4 shows that the RDA-PUF method has a significant reduction
in energy consumption when related to the conventional procedure because it deliberates
the consistency, quantity and less energy established from base to endpoint. Furthermore,
as the number of nodes in the arrangement grows, so does the amount of energy consumed
also grows.
43

Fig. 4. Graphical view of Energy Consumption
6.2 Performance of End-to-End Delay
Figure 5 shows the investigation of end-to-end delay using RDA-PUF and existing ap-
proaches (FDEAM [23], EOMR [24], and TSGWO [25]). Figure 5 clearly illustrates that
RDA-PUF has the shortest delay and is constantly attempting to send data packets to
the best intermediary node, satisfying the ideal characteristic. On the other hand, the
traditional approaches took a long time to determine the endpoint node. Furthermore, it
initiates the prolonged return after the endpoint node. As a result, the associated sys-
tem’s usual delay is longer than the RDA-PUF technique. The investigation of end-to-end
delay is tabulated in Table 3. When compared to existing methodologies, the suggested
RDA-PUF achieves a shorter time delay, as shown in table 3.
Table 3. Investigation of End-to-end delay
Number of Nodes End-to-End Delay Performance
100 8 10 8 7
150 14 17 15 8
200 15 23 19 10
250 23 30 26 16
300 28 35 31 19
350 32 42 37 22
400 40 50 44 25
6.3 Performance of PDR
Figure 6 shows a comparison between RDA-PUF and existing approaches (FDEAM [23],
EOMR [24], and TSGWO [25]) in terms of packet delivery ratio performance. When com-
pared to existing approaches, figure 6 appears to suggest that RDA-PUF yields greater
PDR. Because, it recognizes that when the matching communication is single and binary,
the PDR ratio becomes superior, and the overall assessment remains the same. The com-
parative examination of Packet Delivery Ratio performance is tabulated in Table 4. When
compared to existing approaches, the proposed RDA-PUF, transfers additional packets as
observed in table 4.
6.4 Performance of Network Lifetime
Figure 7 shows comparative analysis of network lifetime amongst the proposed RDA-PUF
and the conventional methodologies. In RDA-PUF, there is a longer lifespan than existing
44

Fig. 5. Performance of End to End Delay
Table 4. Performance of PDR
Number of Nodes Performance of PDR
100 80 70 73 84
150 86 84 85 90
200 86 85 87 92
250 90 87 89 93
300 92 88 89 94
350 92 88 90 95
400 94 89 92 97
FDEAM [23], EOMR [24], and TSGWO [25], because as the node count rises, extra nodes
begin steering packets indiscriminately, and there is a higher chance that a node may
expire at some point. Only the optimal node is assigned to transfer packets in the RDA-
PUF approach, resulting in increased battery life and network longevity. The comparative
examination of Packet Delivery Ratio performance is tabulated in Table 5. It displays that,
when compared to conventional methods, the suggested RDA-PUF increases the network
lifetime.
Table 5. Comparison of Network Lifetime
Number of Nodes Performance of Network Lifetime
FDEAM[23] EOMR [24] TSGWO[25] Proposed RDA-PUF
50 1245 1150 1220 1330
100 1170 1200 1208 1331
150 1138 1070 1109 1258
200 1072 990 1010 1127
250 1007 920 994 1072
300 968 860 951 982
350 927 853 915 985
From the analysis of all the performance parameters; Energy Consumption, End-to-
End Delay, PDR and Network Lifetime, it can be concluded that an improved QoS in WSN
is accomplished by multipath routing via RDA-PUF. The proposed RDA-PUF enhances
the network performance via exploiting routing method to move the data from BS to end
point, by selecting the best path.
45

Fig. 6. Performance of PDR
Fig. 7. Performance of Network Lifetime
6.5 Performance of Secure Connectivity
The possibility that two individuals can generate a session key to secure communications
is characterized as a network’s security connectivity. By considering ”intra-cluster secure
connectivity” a bilateral password is created in the network by a cluster-member, because
this study primarily presents a method for intra-cluster authentication and key distri-
bution. The effectiveness of secure connectivity is shown in Figure 8. The probability of
sharing a key grows as the amount of preloaded keys increases, as seen in Figure 8. As
the key pool size grows larger, the possibility of sharing a key reduces for almost the same
settings.
Table 6. Comparative analysis Secure Connectivity
Key Pool Size Existing PUF-MAKD [21] Proposed RDA-PUF
2000 1.098 1.211
4000 1.097 1.209
6000 1.096 1.206
8000 1.095 1.204
10000 1.094 1.201
6.6 Performance of Time Duration
The ordinary time duration needed to implement associated cryptographic processes on
the hardware tools through the system clock is fixed to 32MHz. And the below figure
46

Fig. 8. Performance of Secure Connectivity
9 displays the time duration performances of existing CL-PKC [28] and proposed RDA-
PUF protocol. From the figure 9, it clearly shows that proposed RDA-PUF achieves mutual
authentication and key exchange in short interval of time. Table 7 shows the comparative
analysis of time duration.
Table 7. Comparative analysis of Time Duration
Rounds Existing CL-PKC [28] Proposed RDA-PUF
1 2.41 2.33
2 2.33 2.27
3 2.36 2.30
4 2.38 2.32
5 2.46 2.35
6 2.35 2.27
7 2.31 2.24
8 2.42 2.29
9 2.49 2.33
10 2.43 2.30
Fig. 9. Performance of Time Duration
6.7 Data Integrity and Verification
The detected data is encrypted using a shared symmetric key in this technique. The com-
bined result is sent to the sink, where it is double-checked. For the sensed findings, this
47

procedure is iterated by every sensor. Figure 10 shows the data integrity performance
through detection ratio. From the figure 10, it clearly shows that proposed RDA-PUF
achieves higher detection ratio when compared to existing Data Validation and Integrity
Verification for Trust Based Data Aggregation (DVIVTDAP) [32]. By examining the sim-
ulation results, it is indicates that the best method improves data accuracy.
Fig. 10. Performance of Detection Ratio
7 Conclusion
In recent trends, WSNs have attracted a lot of attention as it has high resource constraints
and principal concerns in computer technology. When transmitting point-to-point, how-
ever, these technologies have limited bandwidth, consumption, and resources. In sensor
networks, how crucial information may be analyzed in a more energy-efficient approach,
is a hot topic. In this research, RDA-PUF is offered as a method for transporting data
from IoT to a target node with a longer delay. The proposed RDA-PUF could greatly
improve secure communication in context of appropriate RSA protocol. Using a PUF in
place of key administration and verification element to ensure protection in contradiction
of malicious attacks. The overall simulation shows that the proposed RDA-PUF produced
better results by reducing energy usage, end-to-end latency, network lifetime and PDR of
proposed RDA-PUF is improved by 37%, 12%, 6.3% and 3.09%, respectively when related
with existing FDEAM, EOMR and TSGWO technique. Similarly, the projected RDA-PUF
enables highly secure connectivity (1.211) than the existing PUF-MAKD (1.098). In up-
coming work, this technology can be broadened to include novel encryption methodologies
for improving network performance.
References
1. Shen, Jian, Anxi Wang, Chen Wang, Patrick CK Hung, and Chin-Feng Lai, An efficient centroid-based
routing protocol for energy management in WSN-assisted IoT, IEEE Access, Vol. 5, pp. 18469-18479,
2017.
2. Shrivastav, Khyati, and Kishore D. Kulat, Energy efficient scalability of three level hexagonal heteroge-
neous broad transmission distance protocol (3L-HEXA-HTBTDP) for WSN-IoT networks, International
Journal of Communication Systems, Vol. 31, no. 17, pp. e3809, 2018.
3. L. Sasirega, Trust Establishment For Detecting Aggressor Nodes And Improving Route Stability In
Wsn-Iot, Turkish Journal of Computer and Mathematics Education (TURCOMAT), Vol. 12, No. 11,
pp6012-6020, 2021.
4. Reddy, M. Praveen Kumar, and M. RajasekharaBabu, A hybrid cluster head selection model for Internet
of Things”, Cluster Computing, Vol. 22, no. 6, pp. 13095-13107, 2019.
5. Jain, Bindiya, GursewakBrar, Jyoteesh Malhotra, and Shalli Rani, A novel approach for smart cities in
convergence to wireless sensor networks, Sustainable cities and society, Vol. 35, pp. 440-448, 2017.
48

6. Hussein, Walaa, Jiwa Abdullah, and N. A. M. Alduais, Data Aggregation Algorithms with Multiple
Sensors in Clustered-Based WSN/IoT, International Journal of Computing and Digital Systems, Vol.
9, No. 03, 2020.
7. Ullah, A., Azeem, M., Ashraf, H., Jhanjhi, N.Z., Nkenyereye, L. and M. Humayun, Secure Critical Data
Reclamation Scheme for Isolated Clusters in IoT enabled WSN, IEEE Internet of Things Journal, 2021.
8. Reddy, M. Praveen Kumar, and M. RajasekharaBabu, Implementing self adaptiveness in whale opti-
mization for cluster head section in Internet of Things, Cluster Computing, Vol. 22, No. 1, pp. 1361-1372,
2019.
9. Hao, Wang Zhi, Gwo-JiunHorng, and Gwo-Jia Jong, A New Bio-Inspired for Cooperative Data Trans-
mission of IoT, IEEE Access, Vol. 8, pp. 161884-161893, 2020.
10. Anandkumar, M. (2020) “Multicast routing in WSN using bat algorithm with genetic operators for
IoT applications”, Journal of Networking and Communication Systems, Vol. 3, No. 2.
11. El-Sayed, Hamdy H, and Hilal Al Bayatti, (2021) “Comparisons of Some Multi-Hop Routing Protocols
in Wireless Sensor Networks”, Energy, Vol. 5.
12. Aadri, Alaa, and Najlae Idrissi, (2017) “A Cluster based routing algorithm minimizing energy con-
sumption in WSN”, International Journal of Wireless Mobile Networks (IJWMN), Vol. 9, pp49-62.
13. Ara, Tabassum, M. Prabhkar, and PritamGajkumar Shah. (2018) “Energy efficient secured cluster
based distributed fault diagnosis protocol for IoT”, International Journal of Communication Networks
and Information Security Vol. 10, No. 3, 539.
14. Dwivedi, Anshu Kumar, Awadhesh Kumar Sharma, and Pawan Singh Mehra. (2020), “Energy efficient
sensor node deployment scheme for two stage routing protocol of wireless sensor networks assisted iot”,
ECTI Transactions on Electrical Engineering, Electronics, and Communications Vol. 18, No. 2, pp158-
169.
15. L. Xu, G. MP O’Hare, and R. Collier, A smart and balanced energy-efficient multihop clustering
algorithm (smart-beem) for mimoiot systems in future networks, Sensors, Vol. 17, No. 7, 1574, 2017.
16. L. Kaur, and R. Kaur, Comprehensive Survey for Energy-Efficient Routing Protocols on WSNs Based
IoT Applications Enhancing Fog Computing Paradigm, In Applications of Artificial Intelligence in
Engineering, pp. 339-354, 2021.
17. G. Sathishkumar, Bivariate Regression Adaptive Wald’s Boost Energy Aware Routing for Wsn with
IoT, Turkish Journal of Computer and Mathematics Education (TURCOMAT), Vol. 12, No. 7, pp2224-
2241, 2021.
18. A.J. Manuel, G. Gopal Deverajan, R. Patan, Amir H. Gandomi, Optimization of routing-based clus-
tering approaches in wireless sensor network: Review and open research issues, Electronics, Vol. 9, No.
10, pp. 1630, 2020.
19. H.K. Sekhon, and V. Singh, Clustering in Wireless Sensor Network: A Review”, International Journal
of Computers Technology, Vol. 16, No. 6, 2017.
20. J. Marietta, and B.C. Mohan, A review on routing in internet of things, Wireless Personal Communi-
cations, Vol. 111, No. 1, pp. 209-233, 2020.
21. Y. Liu, Y. Cui, L. Harn, Z. Zhang, H. Yan, Y. Cheng, and S. Qiu, PUF-Based Mutual-Authenticated
Key Distribution for Dynamic Sensor Networks, Security and Communication Networks, 2021.
22. M.A. Alharbi, M. Kolberg, and M. Zeeshan, Towards improved clustering and routing protocol for
wireless sensor networks”, EURASIP Journal on Wireless Communications and Networking, Vol. 2021,
No. 1, pp. 1-31, 2021.
23. I. Abasıkeleş-Turgut, and G. Altan, A fully distributed energy-aware multi-level clustering and routing
for WSN-based IoT, Transactions on Emerging Telecommunications Technologies, pp. e4355, 2019.
24. K. Jaiswal, and V. Anand, EOMR: An energy-efficient optimal multi-path routing protocol to improve
QoS in wireless sensor network for IoT applications, Wireless Personal Communications, Vol. 111, No.
4, pp. 2493-2515, 2020.
25. N. Chouhan, and S. C. Jain, Tunicate swarm Grey Wolf optimization for multi-path routing protocol
in IoT assisted WSN networks, Journal of Ambient Intelligence and Humanized Computing, pp. 1-17,
2020.
26. V. Shilpa, and A. Vidya, A Hybrid Optimization Algorithm and Shamir Secret Sharing Based Secure
Data Transmission for IoT based WSN, International Journal of Intelligent Engineering and Systems,
Vol. 14, No. 6, pp. 498-506, 2021.
27. V. Shilpa, A. Vidya, and S. Pattar, MQTT based Secure Transport Layer Communication for Mutual
Authentication in IoT Network, in: Proceedings of Global Transitions Proceedings, In Press 2022.
28. S. Li, T. Zhang, B. Yu, and K. He, A provably secure and practical PUF-based end-to-end mutual
authentication and key exchange protocol for IoT, IEEE Sensors Journal, Vol. 21, No. 4, pp. 5487-5501,
2020.
29. Z. Zhang, Y. Liu, Q. Zuo, L. Harn, S. Qiu, and Y. Cheng, PUF-based key distribution in wireless
sensor networks, Computers, Materials Continua, Vol. 64, No. 2, pp. 1261-1280, 2020.
49

30. Del-Valle-Soto, Carolina, Carlos Mex-Perera, Juan Arturo Nolazco-Flores, Ramiro Velázquez, and
Alberto Rossa-Sierra. ”Wireless sensor network energy model and its use in the optimization of routing
protocols, Energies, Vol. 13, No. 3, pp. 728, 2020.
31. A.M. Fathollahi-Fard, M. Hajiaghaei-Keshteli, and R. Tavakkoli-Moghaddam, Red deer algorithm
(RDA): a new nature-inspired meta-heuristic, Soft Computing, Vol. 24, No. 19, pp. 14637-14665, 2020.
32. S.E. Roslin, Data validation and integrity verification for trust based data aggregation protocol in
WSN, Microprocessors and Microsystems, Vol. 80, pp. 103354, 2021.
Authors
Shilpa Venkata Rao is currently working as Assistant Professor in Department of Com-
puter Science and Engineering at Cambridge Institute of Technology, Bangalore, India.
She received her Bachelor of Engineering degree in Information Science and Engineering
from Visvesvaraya Technological University, Belagavi and Master of Engineering degree
in Computer Networking from University Visvesvaraya College of Engineering, Bangalore
University, Bengaluru. She is currently pursuing Ph.D. under the guidance of Dr. Vidya A
in Computer Science and Engineering at Vivekananda Institute of Technology, affiliated to
Visvesvaraya Technological University, Bangalore, India. Her research interest includes In-
ternet of Things, with a focus on security and authentication protocols, Cloud Computing.
Vidya Ananth is currently working as Professor and Head of the Department of Com-
pute Science and Engineering at Vivekananda Institute of Technology, Bangalore, India.
She received her Bachelor of Engineering and Masters of Engineering degrees in Computer
Science and Engineering from Bangalore University, Bangalore. She was awarded Ph.D.
in Computer Science and Engineering from Jawaharlal Nehru Technological University,
Hyderabad. Her area of research includes Data Mining, Soft Computing, Pattern Recog-
nition, Computer Networks and Image Processing. She is a Life Member of Indian Society
for Technical Education, Computer Society of India and International Association of En-
gineers. She is a Fellow of Institute of Engineers.
50

Hybrid Red Deer Algorithm with Physical Unclonable Function for Security Enhancement in IoT-WSN

More Related Content

Similar to Hybrid Red Deer Algorithm with Physical Unclonable Function for Security Enhancement in IoT-WSN (20)

More from IJCNCJournal (20)

Recently uploaded (20)

Hybrid Red Deer Algorithm with Physical Unclonable Function for Security Enhancement in IoT-WSN