A Blockchain-based Approach for Optimal and Secure Routing in

Wireless Sensor Networks

Hilmi Lazrag

, Rachid Saadane

and Moulay Dirss Rahmani

LRIT Lab, FSR, Mohammed 5 University, Rabat, Morocco

EHTP, Km 7 Route El Jadida, Oasis, Morocco

Keywords:

WSN; Blockchain; Trafﬁc Load; SINR; Security; Digital Signature

Abstract:

The trafﬁc load balance, the interferences reduction, and the security during the routing phase in wireless

sensor network (WSN) are investigated in this paper. In our work, we suppose that the network’s nodes are

sensing some events which generate heavy data that must be carried over several packets. We propose a routing

protocol which takes use of the Blockchain technology to offer a shared memory between network’s nodes.

These nodes are considered as coins which the ownership transacts between the source nodes and the sink. All

the transactions are stored in the Blockchain as a mean to share the network’s status, in real time. In order to

select the optimal path, we introduce a cost function which considers the load density and interferences level

at each node. Furthermore, we are taking advantage of the Blockchain security to secure the selected paths in

the network. The simulation results have shown that this solution could be applicable and could resolve the

issues cited above.

1 INTRODUCTION

Nowadays, wireless sensor networks (WSN) are be-

ing wildly used in several domains (Akyildiz et al.,

2002), due to their simple implementation and,

continuously, improved features (Callaway, 2004),

(Limin et al., 2005). This technology gained more

importance in the IoT era and had been well investi-

gated. However, WSNs are confronted to many prob-

lems. In this paper we are interested in issues related

to the trafﬁc load, the interference, and the security in

such networks during the routing phase. Indeed, as

any distributed wireless system, this type of networks

suffer from the absence of a global state and a global

clock, which makes it very difﬁcult to predict which

path is the most suitable to avoid the trafﬁc load un-

balance and to reduce the interferences levels. Fur-

thermore, the WSNs are usually deployed outdoor, in

harsh and hostile environments, which makes it hard

to maintain high security levels.

In this paper we will introduce the concept of

Blockchain introduced by (Satoshi, 2008), which has

proved its efﬁciency in the cryptocurrency and other

distributed systems, in order to face the issues cited

above. At the ﬁrst stage of our approach, we consider

the network’s nodes as coins, all owned by the sink

(i.e, the nodes owned by the sink are considered as

inactive). Each time a node desires to send a mes-

sage, it calculates the best path to transmit on. Then it

asks the sink to transmit the path’s nodes ownership to

it. The transaction is stored in the Blockchain and the

path’s nodes becomes active. In fact, we could check

for nodes activity directly in the Blockchain. Though,

it seems simple, the path choice requires some ad-

vanced logic. Actually, the transmitting node has to

ﬁnd the optimal path based on the trafﬁc load over

all the network’s nodes and the interference generated

by each node that is transmitting in the same time.

So, we introduce, also, some cost function, which

will help the nodes in selecting the routes. The se-

curity of the transaction is supposed to be granted by

the blockchain’s signatures. In order to validate this

approach, we implemented a lightweight blockchain

mechanism and we used it to simulate our proposition

through an homogeneous static network.

To our best knowledge, our work is the ﬁrst WSN

routing mechanism which considers the Blockchain

as a support to share network status in real time in

order to enhance the routing process. The main con-

tributions of this paper are as follows:

• Considering nodes as coins and transfer their

ownership between each other;

• Use Blockchain as a shared memory to broadcast

Lazrag, H., Saadane, R. and Rahmani, M.

A Blockchain-based Approach for Optimal and Secure Routing in Wireless Sensor Networks.

In Proceedings of the 1st International Conference of Computer Science and Renewable Energies (ICCSRE 2018), pages 175-180

ISBN: 978-989-758-431-2

175

the status of the network’s nodes;

• Use the past nodes’ activities to determine the

trafﬁc load.

The paper is organized as follows : Section 2

presents the Blockchain technology breﬂy. Section 3

describes how we attend to introduce the Blockchain

in the routing phase and how we use the stored trans-

actions to optimize the routing. Simulation results are

illustrated in Section 3. Finally, in Section 4, we give

a brief conclusion for this work.

2 THE BLOCKCHAIN

Blockchain appeared, originally, with the crypto-

currency ’Bitcoin’ in 2008. Since then Blockchain

has gained a strong reputation and had been used in

several domains ranging from economy (Swan, 2015)

to property acquisition (Peir

o and Martinez, 2017)

and health care (Kuo et al., 2017).

Blockchain is considered as a distributed ledger

of transactions, which plays the role of a database

that could be shared over a network of various peers

(UK Government, 2015). Blockchain technology

was designed to face the digital currencies ownership

and the transactions settlement challenges (Satoshi,

2008). It provides, also, a secure mechanism for elec-

tronic collaboration which is accomplished without

a central trusted party (Satoshi, 2008), (UK Govern-

ment, 2015). When a transaction from peer A to peer

B is made, the information is shared with all the net-

work’s peers, underlying the Blockchain. The trans-

action’s credentials are combined in a block, marked

with a timestamp and added as a new block to the

blockchain. Since all transactions use public-private

key cryptography, each transaction could be veriﬁed

and conﬁrmed by all involved parties. The blocks are

appended consecutively, which allows a transparent

and total knowledge of all transactions made in the

past (Dwyer, 2015), (B

ohme et al., 2015).

What one shall retain, is that Blockchain offers a

distributed database which keeps track of all trans-

actions made over a network in a secure way. For

a better understanding of the Blockchain technology

mechanism we refer the reader to (Satoshi, 2008),

(Swan, 2015) and (Franco, 2015).

3 SYSTEM MODEL

We consider a static network consisting of n nodes

and one sink as shown in Figure 1. Each node is sup-

posed to have a map of the network (number of nodes,

their ids, their locations as well as their connectivi-

ties). we assume that the nodes are homogeneous in

term of transmission power and bandwidth. the sink is

supposed to have enough bandwidth and resources to

handle all incoming messages. The sensed events ar-

rive randomly at each node. we assume that the events

generate some heavy data that will be carried through

several packets. These packets have the same size and

they are concidered as a single message. Furthermore,

we assume that the nodes and the sink have an unlim-

ited source of energy (photo-voltaic ...). So, the rout-

ing services could be discussed without being bound

in the fetters of the energy shortage constraints.

Figure 1: The simulated network: the (gray) dotes represent

the sensor nodes and the (red) square represents the Sink.

Even though each node has a good knowledge of

the network’s distribution, it still can’t know the state

of the network at a given time. For example, at an

instant t a node couldn’t tell which nodes are trans-

mitting and which are not. This ignorance is due to

the absence of a shared memory. Additionally, a node

could not predicts whether its peers have been com-

promised or not. Hence, the nodes could, naively,

send/receive the data to/from some malicious peer,

which could offense the integrity and the sanity of the

carried data.

In order to make the network’s state shareable, we

introduce the blockchain mechanism cited above.

3.1 Blockchain as Shared Memory in

WSN

As we said before, the Blockchain system is based on

a ledger which keeps track of all the transactions cir-

culating in a network. Thus, as we need some way to

ﬁgure out which nodes are transmitting and through

which path, we will store the paths which are active,

in real time, as transactions in the Blockchain. To

achieve this, we treat the network’s nodes as coins.

ICCSRE 2018 - International Conference of Computer Science and Renewable Energies

176

More precisely, when some nodes are carrying a mes-

sage from a source node to the sink, their ownership

will be affected to the source node. At the ﬁrst stage

all the nodes are owned by the sink. Each node that is

owned by the sink is considered to be inactive. Oth-

erwise, all the nodes which are not owned by the sink

are considered as active. When a node senses some

event, it looks up in the Blockchain and deﬁnes a list

of all inactive nodes, then it ﬁnds among them which

ones optimize its path to the sink, we will describe

the route choice process in the next subsection. Next,

it asks the sink to transfer the path’s nodes owner-

ship to it. Once the transaction is registered to the

Blockchain the node starts transmitting over the cho-

sen route. When the data is carried successfully to the

sink the transmitting node transfers back the owner-

ship of the path’s nodes, including itself, to the sink

as a way to inform the network’s peers that the trans-

mission was ﬁnished and these nodes were released.

We assume that a source node could own u nodes

while u ≤ n. We assume also that the nodes trans-

mit over two channels, the ﬁrst one is dedicated to

the paths claiming and to the Blockchain transactions

transferring, and the second one is designated to carry

the sensed data. We are interested, primarily, in the

second channel which is used to transmit the message.

We suppose, also, that each intermediate node could

be owned, only, by one source node and a source node

is owned, only, by itself. When a node senses an event

while it is owned by an other node, the latter waits

until its ownership is transfered to the sink. In the

meantime, the node notiﬁes the sink, through the ﬁrst

channel, in order to be added to a waiting queue. The

waiting queue is mainly managed by the sink and it is

necessary to apply, a kind of, priorities to the waiting

nodes.

As we have seen, this technique allows for a good

knowledge of the source nodes as well as the paths

which they transmit on, at a given moment. It is

necessary, also, to mention that the nodes are rep-

resented in the Blockchain by their Ids. Hence, the

trafﬁc load could be, easily, determined through the

Blockchain. Actually, it sufﬁces to determine, di-

rectly from the chain, how many times the status of

a node has changed to be active. This changes num-

ber is, obviously, the number of messages carried by

a node, since a node status changes only when it is

in the path on which a message is transmitted. Now,

after we deﬁned the trafﬁc load at each node, we

have to deﬁne the routing determination process of

our model.

3.2 Route Determination Process

As each node knows the network’s map and as each

one can access the Blockchain and ﬁnd which nodes

are transmitting and which nodes are not, it becomes

simpler to deﬁne the shortest path to the sink through

a set of inactive nodes. However, as said previously,

our main goal is to balance the trafﬁc load and to re-

duce the interferences in the routing phase. So, we

have to deﬁne a cost function which optimizes the

path. First of all let us deﬁne the signal and inter-

ference to noise ratio (SINR) as (Gupta and Kumar,

2000),

SINR

(i, j)

i, j







∑

k=1

k6=i

k, j







(1)

where p

is the transmission power of the i

node,

i, j

is the distance between two nodes i and j, a is the

path loss exponent, and N

is the power of an additive

white Gaussian noise. The equation (1) is used beside

the load trafﬁc to determine the routing cost to the

next hop. The cost function is deﬁned as follow,

Cost

= SINR

(i, j)

/(1 + θ

) (2)

where j is the index of the next hop, SINR

(i, j)

is the

signal to interference and noise ratio, and θ

is the

trafﬁc load of the j

node.

When an event is detected and a message is ready

to be sent, the source node k starts listing all the in-

active nodes, as explained before. Next, it calculates

the routing cost, using equation (2), for each of the

inactive nodes and determines the optimal path us-

ing dijkstra’s algorithm (Dijkstra, 1959). Once the

path determined, a chain veriﬁcation is applied to

each node of the chosen path. If the chain of all the

nodes is veriﬁed, the source k claims for the owner-

ship of these nodes and the transaction is registered to

the Blockchain. Otherwise, k discards the untrusted

nodes and redeﬁnes a new optimal path. In case no

valid path is found to reach the sink, the source node

waits for active nodes to be libereted and notiﬁes the

sink, through the ﬁrst channel, in order to be added to

the waiting queue.

3.3 Security in the Routing Phase

We stated in the previous subsection that each source

node accomplishes a chain veriﬁcation, right after it

chooses the optimal path, for each node on the cho-

sen route and we highlighted that all untrusted nodes

A Blockchain-based Approach for Optimal and Secure Routing in Wireless Sensor Networks

177

(those which has not been veriﬁed) are discarded. Ac-

tually, this kind of ﬁltration guaranties that each se-

lected node, that ﬁgures on any path linking all source

nodes to the sink, have, necessarily, a clean connexion

history, since the nodes that had been on the path of

some malicious peer are ignored. Indeed, it is conve-

nient to say that the routing phase has a security layer

which lies on the Blockchain’s security mechanism

shown in (Satoshi, 2008), (Swan, 2015) and (Franco,

2015). Hence, the proposed routing approach has a

quiet strong security, especially, with the use of some,

trusted, cryptographic algorithms such as ECDSA224

(Johnson et al., 2001) and sha512 (Gueron et al.,

2011).

Overall, the proposed routing approach shows

that, despite the fact that the process of ﬁnding paths

is done through some, fairly, complex calculations, it

is still secure and optimal, involving multiple crypto-

graphic techniques.

4 RESULTS

To validate our work we simulated a network of 55

nodes which contains one sink. The network is a con-

nected graph and distributed as shown in Figure 1.

The simulation was done using Python3, numpy and

matplotlib. First of all we created a basic Blockchain

mechanism, which responds to the requirements of

our work, in term of transactions structure and cryp-

tographic mechanisms. We assumed that the nodes

are broadcasting at a transmission power of 1mW and

we considered that the path loss and the noise are 2

and 5 ∗ 10

−15

, respectively. Next, in order to verify

the ability of the proposed mechanism to balancing

the trafﬁc load, we assumed that only one node is

transmitting , several messages, over the network. To

evaluate whether there is an enhancement or not, we

applied a shortest path routing protocol to the same

scenario and we analyzed the returned data.

Figure 2 shows a comparison of the trafﬁc load,

when only one node is transmitting, using a shortest

path routing protocol against our proposed approach.

As we can see in Figure 2 (a.1), when a shortest

path routing protocol is used, the load density is con-

centrated at a tiny straight region, the brightest area on

the illustration (i.e. the brighter a region is the higher

the trafﬁc load is). By taking a closer look at Figure

2 (a.2) we could distinguish that the load density is

concentrated only around the shortest path between

the source node (green dot) and the sink (red square).

However, if we take a look at Figure 2 (b.1), when we

use our proposed protocol, we could remark that the

brightness is concentrated at two points and it fades

Figure 2: The trafﬁc load when only one node is transmit-

ting: the (green) dot represents the source node. (a.1) shows

the density of load and (a.2) shows the implied nodes, while

using a shortest path protocol. (b.1) shows the density of

load and (b.2) shows the nodes implied in the transmission,

while using the proposed approach.

while we move away from them. These two points

are the source node and the sink, as we can tell from

Figure 2 (b.2). So, rather than be concentrated around

a single path, the trafﬁc load is spread over the net-

work, which highlights the efﬁciency of our system

to resolve the load balancing issue.

Now, let’s analyze the interferences level for the

shortest path routing protocol and our proposed pro-

tocol. To do so, three source nodes are considered: the

ﬁrst one senses an event at t

, the second one senses an

event at t

, and the last one senses an event at t

. Once,

an event is sensed the implicated node selects a path

and starts transmitting over it. We assume that when

one of, the cited above, source nodes starts sending

data it keep transmitting until the simulation ends.

Figure 3 shows the selected routes, when only one

source node is transmitting, when two source nodes

are transmitting, and when three source nodes are

transmitting, simultaneously, while using, both, the

shortest path routing protocol and our proposed pro-

tocol.

Obviously, when using the shortest path routing

protocol and for the three cases, shown in Figure 3:

(a.1), (b.1), and (c.1), the source nodes select the

shortest paths to the sink. When, using our proposed

protocol, we could, clearly, depict that the path choice

in the ﬁrst case, when only one source node is trans-

mitting, mimics the one chosen by the shortest path

ICCSRE 2018 - International Conference of Computer Science and Renewable Energies

178

Figure 3: The path choices: the (green) path represents the

ﬁrst route, the (cyan) one represents the second route and

the (yellow) represents the third route. The left column is

dedicated to the choices made by a shortest path protocol,

while the right one shows these made by the proposed pro-

tocol.

routing protocol (see Figure 3: (a.2)). This similar-

ity is, totally, comprehensible since the load density

is supposed to be equal over the network’s nodes and

since we are using dijkstra’s algorithm to deﬁne our

paths. However, for the two remained cases we could

see, as shown in Figure 3: (b.2) and (c.2), that our

proposed mechanism chooses longer routes which are

located far from each others. As known, the farther a

transmitting node is from its peers the less the inter-

ferences are and the nearest it is the higher the in-

terferences became, which means that our approach

decreases the interferences between the transmitting

nodes.

Finally, after resolving the trafﬁc load balancing

and the interference issues, let’s run one last simu-

lation where we could evaluate the security of our

proposition. First of all we change manually the own-

ership of some nodes and set it to an unknown owner

directly into the block chain which will mess up the

blocks signature. These nodes will be considered as

corrupted by the system. So, let see how the system

will behave to face this issue.

Figure 4 shows the load density when some of the

network’s nodes are compromised. As shown in Fig-

ure 4 (a.1), there are two areas, on the load density

illustration, that are darker than the remaind regions.

Figure 4: The trafﬁc load when some nodes are com-

promised: the (orange) dotes represent the compromised

nodes. (a.1) shows the load density and (a.2) shows the ex-

act location of the compromised nodes.

The ﬁrst area is located in the bottom of the ﬁgure and

moves diagonally to the up right, and the second one

starts right in top and continues vertically near to the

center. By analyzing the Figure 4 (a.2) we could see

that the dark areas correspond to the orange nodes,

which we intentionally messed up. And as we said

before, the darker an area is the less messages went

through it, so we could deduce that the orange nodes

have been ignored by the benign nodes of the network

during the routes determination processes. Thus, our

protocol has shown that it could offer a layer of secu-

rity in the route determination phase.

5 CONCLUSION

In this paper, optimal routing protocol in wireless

sensor networks based on Blockchain has been dis-

cussed. To achieve a better solution to the trafﬁc load

unbalance, the high interference levels and the secu-

rity issues, we proposed a protocol which takes ad-

vantage of the Blockchain technology beneﬁts. The

approach consists of using Blockchain as a shared

memory between nodes and storing all the network’s

activities on it. The nodes are considered as coins

which are owned by the sink when they are inactive

or owned by the source node if they are carrying some

message. Moreover, a cost function is proposed in or-

der to optimize the chosen path. regarding the proto-

col’s security, it is granted by the Blockchain. Sim-

ulation results show that the proposed protocol can

improve the trafﬁc load balance, decrease the inter-

ference levels and guaranty a strong security during

the routing phase.

A Blockchain-based Approach for Optimal and Secure Routing in Wireless Sensor Networks

179

REFERENCES

Akyildiz, I. F., Su, W., Sankarasubramaniam, Y., and

Cayirci, E. (2002). A survey on sensor networks.

IEEE Communications Magazine, 40:102–114.

ohme, R., Christin, N., Edelman, B., and Moore, T.

(2015). Bitcoin: Economics, technology, and gover-

nance. Journal of Economic Perspectives, 29.

Callaway, E. H. (2004). Wireless sensor networks architec-

tures and protocols. Aerbach publishers.

Dijkstra, E. W. (1959). A note on two problems in connex-

ion with graphs. Numerische Mathematik, 1:269–271.

Dwyer, G. (2015). The economics of bitcoin and similar

private digital currencies. Journal of Financial Stabil-

ity, 17:81–91.

Franco, P. (2015). Understanding Bitcoin: Cryptography,

Engineering, and Economics. Wiley.

Gueron, S., Johnson, S., and Walker, J. (2011). Sha-

512/256. In Proceedings of the 2011 Eighth Interna-

tional Conference on Information Technology: New

Generations, pages 354–358. IEEE Computer Soci-

ety.

Gupta, P. and Kumar, P. R. (2000). The capacity of wireless

networks. IEEE Transactions on Information Theory,

46:388–404.

Johnson, D., Menezes, A., and Vanstone, S. (2001). The

elliptic curve digital signature algorithm (ecdsa). Int.

J. Inf. Secur., pages 36–63.

Kuo, T.-T., Kim, H.-E., and Ohno-Machado, L. (2017).

Blockchain distributed ledger technologies for

biomedical and health care applications. Journal

of the American Medical Informatics Association,

24:1211–1220.

Limin, S., Jianzhong, L., and Yu, C. (2005). Wireless sensor

networks. Beijing: Tsinghua university press, pages

1–20.

Peir

o, N. and Martinez, Garc

ıa, E. (2017). Blockchain and

land registration systems. European Property Law

Journal, pages 296–320.

Satoshi, N. (2008). Bitcoin: A peer-to-peer electronic cash

system. https://bitcoin.org/bitcoin.pdf.

Swan, M. (2015). Blockchain: Blueprint for a New Econ-

omy. O’Reilly Media, Inc., 1st edition.

UK Government, C. S. A. (2015). Distributed ledger tech-

nology: Beyond blockchain. UK Government Ofﬁce

for Science.

ICCSRE 2018 - International Conference of Computer Science and Renewable Energies

180