Research Article

International Journal of Distributed

Sensor Networks
2017, Vol. 13(4)
A distributed data secure transmission Ó The Author(s) 2017
DOI: 10.1177/1550147717705552
scheme in wireless sensor network journals.sagepub.com/home/ijdsn

Wei Liang, Yin Huang, Jianbo Xu and Songyou Xie

Abstract
Sensor nodes around monitoring area have various distances to the target node. Therefore, it is difficult to ensure secu-
rity of broadcasting data transferred from a single wireless sensor node to base station. Multi-hop transmission of data
between sensor nodes wastes network resource. In this case, a distributed data secure transmission scheme is proposed
in a wireless sensor network. Data transmission is classified into two stages: constructing a collection of receiving nodes
and selecting a unique forwarding node from this collection. These are implemented using analysis of relative movement
distance between nodes and transfer time competitive mechanism. Besides, we have assumed a network model for dis-
tributed data secure transmission to improve efficiency of data transfer. This design includes secure model of node com-
petition, data perception model, and anti-resistance model. Moreover, the security of competition transfer for nodes in
wireless sensor network is evaluated. Finally, simulation proves that the proposed scheme has good performance in
security and stability compared to similar schemes.

Keywords
Wireless sensor network, distributed data, distributed data secure transmission, data perception model, security
evaluation

Date received: 14 December 2016; accepted: 23 March 2017

Academic Editor: Fei Yu

Introduction and battlefield. The number of nodes is always

large. It is not practical to change battery or
Wireless sensor network (WSN) is widely applied in charge battery to prolong network lifetime. So,
many fields, such as military battlefield, disaster rescue, low energy consumption is first considered in
environment monitoring, traffic management, trail any technique and protocol to prolong network
report, and medical care.1–3 As shown in Figure 1, WSN lifetime in designing a sensor network.9
consists of mobile sensor nodes and sink node. Sensor (2) Dynamic topology. WSN is a dynamic net-
node is deployed on mobile entity for collecting neces- work.10 The topology structure changes under
sary information. Sink node can be fixed or mobile for the following conditions: (1) node invalidity
receiving information from sensor nodes and transmit- caused by energy, environment, and so on; (2)
ting it to backbone network. Different with traditional addition of new node; and (3) mobile node in
network, this network has features such as rapid mobi-
lity of nodes, dynamical network topology, and require-
ment of high security and reliability.4–8 School of Computer Science and Engineering, Hunan University of
Science and Technology, Xiangtan, China
(1) Limited power capacity. Sensor nodes are bat-
Corresponding author:
tery-powered. But battery capacity always has Wei Liang, School of Computer Science and Engineering, Hunan
limited energy. Besides, sensor nodes are gener- University of Science and Technology, Xiangtan 411201, China.
ally deployed in severe environments like desert Email: [email protected]

Creative Commons CC-BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License
(http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without
further permission provided the original work is attributed as specified on the SAGE and Open Access pages (http://www.uk.sagepub.com/aboutus/
openaccess.htm).
2 International Journal of Distributed Sensor Networks

WSN. DDST divides data transmission into two stages,

namely, constructing a collection of receiving nodes
and selecting a unique forwarding node from this col-
lection. DDST uses relative movement analysis and
competitive mechanism based on forwarding time. This
protocol has the following features: (1) receiver-based
data transmission. Generally, sender-based data trans-
mission needs to create global routing from source
node to sink node. This global routing is not necessary
in receiver-based data transmission. A proper candi-
date node is selected in neighboring nodes of sending
node for receiving data. (2) Success rate of data trans-
mission. This article has set two conditions. First, can-
didate node will not move out of the communication
range in data transmission period. Second, candidate
Figure 1. Structure of wireless sensor network. node should be set within the intersection of communi-
cation area and the 60-degree-sector-oriented sink
node. The candidate should satisfy both conditions. (3)
network. Consequently, WSN can reconstruct Competitive mechanism based on forwarding time.
network dynamically. This mechanism is applied in selection of forwarding
(3) Data centered. Traditional network is connec- node. It forwards data instead of multiple nodes. In
tion centered, but WSN is data centered.11 this case, the number of communication nodes and
When WSN reports monitoring events or users energy consumption are reduced. (4) Secure data for-
inquire events with WSN, only events con- warding protocol. The protocol weeds out candidate
cerned by WSN are reported to users, not con- receiving nodes continuously, changing it from working
cerning the event from which sensor node. This state to dormant state to save energy. In addition, the
feature makes WSN a data-centered network. forwarding time–based competitive mechanism selects
(4) Limited computation ability. Nodes in WSN the forwarding node with the most energy. The simula-
have weaker processor owning to limited vol- tion proves DDST can effectively prolong network life-
ume and power. Complex computation and time and achieve better data transmission rate.
data storage cannot be realized by these nodes. This article is organized as follows. The second part
So, simple algorithm is required in WSN, which gives an overview of related work. The DDST protocol
can be easily implemented in the sensor node.12 is introduced in the third section. After that, the imple-
(5) Self-organization. Sensor nodes are always mentation of DDST is concretely illustrated. The fifth
deployed by aircraft. So, positions and adja- section analyzes and compares the experimental results.
cency relationship of these nodes cannot be Finally, this work is concluded.
determined in advance. Besides, sensor network
has dynamic topology. Self-organization is
required for automatic configuration and man- Related work
agement.13 Multiple-hop wireless network is WSN is designed to sense and acquire information in
automatically constructed by topology-control objective physical world.18–21 Data transmission proto-
mechanism and network protocol. col is the foundation of various applications in mobile
(6) Wide distribution and large scale. Sensor nodes sensor network (MSN) and should be considered in
are always deployed in a wide area, especially network construction. To design a data transmission
in severe environment without human race. protocol, one should consider not only energy con-
Network maintenance is much difficult. To sumption of each node but also network lifetime.
obtain reliable and precious information, large Mobile WSN has different application fields, which
number of sensor nodes will be deployed in the requires various transmission protocols. On this basis,
monitoring area. It reduces or eliminates moni- data transmission protocol in WSN can be classified
toring dead zone caused by node validity. The into four categories, namely, transmission protocols
performance of fault tolerance is enhanced.14 based on replication, clustering, quality of service, and
geographic information. Geographic information–
Nodes in WSN have high energy consumption and based data transmission protocol is widely used. Node
low success rate of data transmission.15–17 We analyze position has large impact on data transmission.22–24 In
previous data transmission protocol and propose a dis- network service, the sensing data of nodes is meaning-
tributed data secure transmission (DDST) protocol in ful only after the geographic information is obtained.
Liang et al. 3

Recently, researchers conducted much work on sta- scheme has balanced transmission efficiency and over-
tic WSN because the mobility of nodes in mobile WSN head. The used theory is simple but effective. High
will change network topology, interrupt, or invalidate transmission efficiency is achieved. Blum et al.32 pro-
the routing path.25,26 So, many data transmission pro- posed an implicit geographic transfer protocol. The for-
tocols in WSN cannot be applied to MSN. warding area is limited within a sector of 60°. Waiting
Some clustering-based data transmission protocols time of sender node is set on basis of the distance
are proposed in MSN. The famous low-energy adaptive between receiver node and sink node and its remaining
clustering hierarchy (LEACH) protocol22 assumes energy. It has solved the problem of multi-cast inhibi-
nodes are static. When these nodes move, LEACH tion and considered energy.
loses lots of data. It proves the protocol is not suitable
for MSN. DS Kim and YJ Chung27 proposed low-
energy adaptive clustering hierarchy–mobile (LEACH- DDST protocol
M) protocol by improving LEACH to support mobile
nodes. In each transmission, an acknowledge message Network model
is sent to cluster head to confirm whether a mobile Data transmission protocol in MSN is studied in this
node can communicate with cluster head. Compared to article. The network model should satisfy four condi-
LEACH, LEACH-M obviously improves data trans- tions as follows:
mission success rate. However, LEACH-M selects clus-
ter head with random probability, which is the same (1) Assume that N sensor nodes with communica-
with LEACH. The mobility of cluster head is not con- tion radius R are randomly deployed in a mon-
sidered. If the cluster head moves out of its cluster itoring area Q, as shown in Figure 2. After
before cluster head rotation, the cluster will be broken, being deployed, these nodes can move with a
causing lots of packet loss. slow speed and each has a unique ID, namely,
Replication-based transmission protocol is that mul- 1, 2, ..., n. The ID of sink node is set as 0. Sink
tiple data copies will be transmitted to improve data node cannot move and other nodes are ran-
transmission success rate. Wang and Wu28 proposed a domly deployed with high density to ensure
message fault tolerance-based adaptive data delivery connectivity of this network.
scheme (FAD). The computation of data transmission (2) Each wireless sensor node can sense its remain-
rate is the same with that of replication-based efficient ing energy in real time. Initial energy of each
data delivery scheme (RED). Queue is managed using node could be different and not be added. This
fault tolerance value. When the queue is full and queue case is more close to real network.
message updates, some messages with the fault toler- (3) Each node in WSN carries a module-like global
ance value greater than the threshold value will be positioning system (GPS). So, the information
dropped. FAD is an improvement of RED, which has such as position, movement speed, and direc-
better data transmission rate. But the energy consump- tion could be sensed by itself.
tion is higher than that of RED. Li et al.29 proposed a (4) The time of all nodes in WSN is synchronized.
receiver-based cross-layer forwarding (RCF) protocol. Data are transmitted by cooperation to avoid
The protocol utilizes a self-adaptive mechanism for for- collision and save energy.
warding right contention. Also, it deals with the data
collision and multi-cast suppression better through the
dual-channel communication model and puts forward
an efficient routing void bypass mechanism. RCF has
good reliability, but dual-channel communication con-
sumes more energy. Lu and Li30 proposed a novel
probability distribution–based dynamic and predicted
(PDDP) data forward strategy. It utilizes mobility of
sensor nodes to create dynamic path which can achieve
a network delay close to optimum value, at the same
time, consume network resource efficiently. This strat-
egy has high robustness and practicality. The geo-
graphic information–based data transmission protocol
can be classified into two categories: sender-based pro-
tocol and receiver-based protocol. Wang and Wu31 pre-
sented an RED. It utilizes history data–based
forwarding method to compute transmission rate and
introduces erasure coding to improve efficiency. This Figure 2. Network model of DDST protocol.
4 International Journal of Distributed Sensor Networks

DDST protocol analysis

This section gives some important definitions before
analyzing DDST protocol:

Definition 1: Communication circle. A node in MSN

moves at any time. The area, used the node as the
center to form a circle with communication radius,
is called communication circle.
Definition 2: Node density. The number of nodes in
unit area at any time is node density, denoted
by ‘‘r.’’
Definition 3: Data transmission success rate. During
a communication period of MSN, the ratio of data
amount received by sink node to data amount sent Figure 3. Role classification of DDST protocol.
by sending node is data transmission success rate.

In this article, a DDST protocol is proposed in small yellow circle. Assume the node of ID 1 is the
WSN. DDST divides data transmission into two stages, sending node with original perception data. The com-
namely, constructing a collection of receiving nodes munication radius is R. The circle with the center of
and selecting a unique forwarding node from this col- node 1 and radius of R is called communication range.
lection. DDST uses relative movement analysis and Node 1 is used to send data. Nodes with ID 11, 12, 13,
competitive mechanism based on forwarding time. This and 14 exceed the communication radius of node 1,
protocol achieves effective, rapid, and precious data which cannot receive data. They can do nothing but
transmission. sleep. The nodes with ID of 6, 7, 8, 9, and 10 are within
The first stage is to construct a collection of receiv- the communication area. They can receive data from
ing nodes. A unique forwarding node is selected from node 1, but cannot be added to collection of candidates
this collection. This article has set two conditions for because they are far away from sink node and cannot
candidates. First, candidate node will not move out of satisfy the second condition. Nodes with ID 3, 5, 15,
the communication range in data transmission period. and 16 are within the communication area, which can
Second, candidate node should be set within the inter- receive data and satisfy the second condition. But the
section of communication area and the 60-degree-sec- first condition is not satisfied. So, these nodes cannot
tor-oriented sink node. The candidates should satisfy be candidates. Node 4 satisfies both conditions, which
both conditions. It reduces data receiving and the will be added in collection of candidates. But the for-
speed, which is energy saving. warding competitive time is long, so it cannot be
It is unnecessary for every node to receive data. selected as forwarding node. Only node 2 satisfies both
After constructing the collection of candidates, the sec- conditions and has shorter forwarding competitive
ond stage starts. This stage selects a unique forwarding time. It is selected as unique forwarding node.
node from the collection of candidates. The competitive DDST is described as follows. At the stage to collect
mechanism based on forwarding time is used in this receiving nodes, communication is unnecessary. When
stage. We introduce this mechanism as follows. After the source node wants to send data, other nodes will
receiving data, the mechanism will compute forwarding compute collection of candidates based on algorithm of
competitive time of each candidate, t = (k  d  T )4Er . receiving nodes. The candidates receive data packages
Here, the parameters Er , d, T , and k are, respectively, from source node. It achieves high data-reception rate
remaining energy, distance to sink node, communica- with lower energy consumption. As shown in Figure 3,
tion period, and regulation parameter. The parameter the sending node 1 broadcasts its location. Nodes with
k makes it satisfy 0\(k  d)4Er \1. The node with the ID from 2 to 16 can receive the information, but nodes
minimum forwarding competitive time is selected as from 11 to 14 are out of communication area. The
data forwarding node. From the expression to compute nodes satisfying both conditions above are suitable for
t, more remaining energy and short distance to sink data transmission. From Figure 3, node 2 and node 4
node make t the minimum. receive data package.
After that, we use competitive mechanism in collec-
tion of receiving nodes to determine node time. First, a
Description of DDST protocol time interval t is set. Each candidate automatically per-
In DDST protocol, data transmission has six cases, as forms forwarding competitive mechanism to compute
shown in Figure 3. Wireless sensor node is denoted by the forwarding competitive time. The node with the
Liang et al. 5

minimum time is selected as the forwarding node. As

shown in Figure 3, the candidates are node 2 and node
4. They locate within the communication area of send-
ing node. When they receive data from node 1, the for-
warding competitive time is computed. Node 2 has
short forwarding time, so it is selected as forwarding
node. Node 4 drops the data package and sleep. After
that, node 2 becomes sending node, repeat above pro-
cedures until the data package from node 1 has been
sent to sink node.

Construction of candidates nodes. The main task of first

stage is to select candidates from neighboring nodes of
sending node. The neighboring nodes analyze their
information and determine whether they satisfy the fol-
lowing two conditions:
Figure 4. Model of relative motion between the nodes.
Condition 1. Candidate node will not move out of
the communication range in data transmission As shown in Figure 4, in the direction parallel to
period. connection of two nodes, relative displacement ||OM||
Condition 2. Candidate node should be set within could be computed when the value of b is different.
the intersection of communication area and the 60- When b is between 0° and 90°, or between 270° and
degree-sector-oriented sink node. 360°

The candidates should satisfy both conditions. For jOM j = d  v1  T  cos a  v2  T  cos b
condition 1, nodes which will not move out of commu- ð1Þ
= d  (v1  cos a + v2  cos b)  T
nication area during period of data sending can be
added to candidate collection. The nodes are mobile, so When b is between 90° and 270°
relative motion analysis between nodes is required to
judge whether it will move out of communication range jOM j = d  v1  T  cos a + v2  T  cos b
during a period. It lays foundation for the second stage. ð2Þ
= d  (v1  cos a  v2  cos b)  T
Only those nodes satisfying conditions can become
receiving nodes. Otherwise, nodes go to sleep.
In the direction vertical to connection of two nodes,
The communication period T of a node is short, and
the relative displacement is computed as follows.
node movement is regarded as uniform linear motion.
When b is between 0° and 180°
We focus on the motion of receiving node to analyze
the relative motion between nodes. The details are illu- jOEj = v1  T  sin a + v2  T  sin b
strated as follows. ð3Þ
The main parameters are set as follows. Assume that = (v1  sin a + v2  sin b)  T
the movement speed of sending node is v1 and candi-
When b is between 180° and 360°
date moves with speed of v2 . The angle between motion
direction of sending node and receiving node is denoted
jOEj = v1  T  sin a  v2  T  sin b
as a. The angle between motion direction of receiving ð4Þ
node and sending node is denoted as b. The communi- = ðv1  sin a  v2  sin bÞ  T
cation radius is R. d and t are distances from sending
node to candidate node and communication period, After time T, the distance d 0 between two nodes
respectively. could be computed through equations (1)–(4)
In Figure 4, node S is the sender and N is a neighbor- qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ing candidate of S. After a period, we need to analyze d0 = jOEj2 + jOM j2 ð5Þ
whether N will move out of communication area of the
sender. After time T, S moves to E and N moves to M. By comparing d 0 to R, we can determine which node
Now, we compute d 0 , namely, the value of ||ME||. will be data receiving node. If d 0 \R, the first condition
d 0 \R demonstrates that N will not move out of com- to be forwarding node is satisfied. Otherwise, it cannot
munication area after time T. The first condition is sat- be the forwarding node. Candidate node should be set
isfied and it can be selected as data receiving node. within the intersection of communication area and the
6 International Journal of Distributed Sensor Networks

Figure 5. Position relationship between candidate nodes and

sink node.

60-degree-sector-oriented sink node as shown in

Figure 5.
In Figure 5, S is sending node, R is communication
radius (m) of S. S has many neighboring nodes. Among
these nodes, A, B, C, D, E, and F can be regarded as
candidates. Although H, I, J, and K are within commu-
nication area, they are not in the interaction area of
communication area and the 60-degree-sector-oriented
sink node. So, they cannot be added to the collection
of candidates. Other nodes can satisfy the second con-
dition of candidates.
Figure 6. Algorithm flowchart of DDST programming.
Selection a unique forwarding node from candidate nodes. In
this section, the forwarding node will be selected from of data forwarding. From equation (6), the forwarding
the candidates using competitive mechanism, namely, competitive time is inversely proportional to the dis-
forwarding-based competitive mechanism. After receiv- tance to sink node and is proportional to remaining
ing nodes receive data, the competition for data for- energy. The node with more remaining energy and
warding starts. Senor nodes could be aware of its short distance to sink node has more chance to be the
information. So, it can compute the forwarding com- unique forwarding node.
petitive time. After that, each receiving node detects the The competition procedure for receiving node to be
busy signal. If there is no busy signal when its forward- forwarding node is described as follows. Each receiving
ing competitive time comes, the node will send a data node can compute its forwarding competitive time.
package to sending node and tell that it can forward Before the competitive time, receiving node first detects
data. When the sending node receives package, inhibi- busy signal. If busy signal is not detected, receiving
tion package is generated to stop sending other request node will send a request frame. The sending node
packages. If busy signal is detected, the received data receives request frame and sends a busy signal to stop
will be dropped. As known, the receiving node with the other receiving nodes sending request frames. In this
minimum forwarding competitive time has superiority case, only a receiving node can be the node to forward
of information interaction with sending node. It suc- data, and other nodes will drop received data.
cessfully becomes the node to forward data.
Two parameters are required to compute forwarding
competitive time: remaining energy Er and distance to Implementation of DDST scheme
sink node d. T is communication period and l is weight
The DDST protocol includes two critical code seg-
value
ments: receiving node–based protocol and forwarding
d competitive mechanism–based protocol. The main flow
Time = l  T ð6Þ of protocol design is shown in Figure 6. Nodes in net-
Er
work are initialized. When the communication period
The consideration of energy and distance to sink comes and communication channel is idle, the first
node is to balance energy consumption and reduce hops thing is to judge the role of node. If the node is a
Liang et al. 7

sending node, it will broadcast its information to candi- Table 1. Simulation parameters.
dates and change its role. Candidates receive this infor-
mation and compute the forwarding competitive time Parameters Values
to select a unique forwarding node. After that, the Network size (m2) 300 3 300
information will be further forwarded to sink node. Number of nodes 200
The proposed algorithm is described as follows. All Movement speed of nodes (m/s) 0.6
nodes are synchronized before starting the algorithm. Position of sink node (120,120)
First, sensor node senses data from neighboring nodes. Initial energy of nodes (J) 100
Communication radius (m) 20–80
On basis of received data and its own information, the Energy consumption in sending (J) 2 3 1023
node can judge whether it can be the receiving node. If Energy consumption in receiving (J) 1 3 1024
both conditions are satisfied, the node will enter the Packet size (bytes) 128
competition of forwarding node. Otherwise, the node Storage space (bytes) 2048
goes to sleep.
After that, random selection algorithm for forward-
ing node is performed. Based on the conditions of
being receiving node, the forwarding node is randomly
selected from receiving nodes. With the forwarding
competitive time, forwarding node is selected for data
forwarding. To address random data transmission, a
distributed data transmission scheme is designed in
WSN. This scheme uses multiple-path competition
optimization between sending node and receiving node.
Encoded data is transformed redundantly. After receiv-
ing enough redundant data, receiving node could
decode it to get original data. When decoding fails,
extra data transmission is necessary. This scheme could
save communication overhead and have good security.

Experimental analysis and comparison

Simulation parameters Figure 7. Analysis of the nodes life cycle.
In this section, simulation experiment is conducted to
verify performance of DDST. Additionally, we analyze be described by the relationship between surviving
node number within sensing area and the effect of vari- nodes and running rounds of network.
ous constraint values on data transmission. The experi- Figure 7 compares network lifetime for three data
ment platform is implemented by C++ language.
transmission schemes, namely, RCF, FAD, and DDST.
Network simulator (NS2) is used to quantify perfor-
We simulate and analyze energy consumption of three
mance in real environment. The nodes are randomly
schemes to evaluate network lifetime. In Figure 7, the
distributed in the target area. Here, the following para-
number of nodes is set as 100. We conducted 850
meters are considered: network size is set as
rounds of experiments. The results show that DDST
300 3 300 m2. The number of nodes in experiment
obviously improves surviving time of nodes by compar-
network is 200. On basis of network model, sensor
ing to RCF and FAD. Although the time of first dead
nodes slowly move at a constant speed. In simulation,
node in DDST is earlier than that of RCF, the node
we set the speed of nodes to be 0.6 m/s. Sink node is
dies with a slow speed rate. Finally, the lifetime of
fixed at position (120,120). Initial energy of each sensor
DDST is longer than that of FAD and RCF. Because
node is 100 J. The communication radius ranges from
DDST uses data transmission based on receiving node,
20 to 80 m. One time of data sending and data receiv-
it is unnecessary to create global routing. Candidates
ing consumes energy 2 3 10023 and 1 3 1024 J,
are selected from neighboring nodes of sending node to
respectively. The size of data package is 128 bytes.
receive data. In addition, a competitive mechanism
Each node can store 2048 bytes of data. All the para-
based on forwarding time is proposed to select a unique
meters are set as shown in Table 1.
forwarding node. The forwarding node transmits data
instead of multiple nodes. It reduces energy consump-
Comparison of network lifetime tion. So, DDST is an energy-efficient protocol. In
Network lifetime is a critical indicator to evaluate per- design of this protocol, it weeds out the candidate
formance of data transmission protocol in MSN. It can receiving nodes continuously, changing it from working
8 International Journal of Distributed Sensor Networks

Figure 8. The network lifetime under energy-constrained Figure 9. Comparison of transmission success ratio with
environment. various communication radii.

state to dormant state to save energy. In addition, the

forwarding time–based competitive mechanism makes
the forwarding node be that of more energy. In this
case, network lifetime is prolonged.
From Figure 8, all data transmission algorithms
(RCF, FAD, and DDST) have the same rate to gener-
ate message. Some algorithms consume energy rapidly,
so less data is transmitted. By comparing to above
simulation results, DDST can obviously prolong net-
work lifetime and have better performance of data
transmission under energy constrained environment.
Under the same WSN environment, different trans-
mission methods lead to different network lifetimes. As
shown in Figure 8, under energy-constrained environ-
ment, DDST has long network lifetime and better secu-
rity compared to other transmission methods because
DDST is a dynamic distributed data transmission with Figure 10. Comparison of transmission success ratio with
less energy consumption. Compared to other methods, various sensor nodes.
it consumes less energy and time in data transmission
and prolongs network lifetime.
transmission success rates. The experiment results are
shown in Figure 9. Communication radius ranges from
Comparison of data transmission performance 10 to 60 m. Obviously, when communication radius
The authors compare the DDST scheme with several ranges from 10 to 30 m, the hops from a node to sink
classic routing algorithms (RCF and FAD) in terms of node are increasing in three algorithms. DDST has set
routing performance and energy efficiency. We mainly two conditions in selection of candidates. Candidates
consider three indicators including transmission success should satisfy both conditions, which makes the hops
ratio, delivery cost, and delivery delay. Transmission of DDST to sink node be less than that of other two
success ratio is the ratio of the number of successful methods. With the increase in communication radius,
transmitted messages from source node to target node the superiority of DDST is not obvious.
to the number of totally generated messages. Delivery We also compare the transmission performance with
cost is the energy consumption during data transmis- various number of sensor nodes. The simulation and
sion. Delivery delay is the time interval of the message comparison are illustrated as follows. Figure 10 shows
generated from source node and sent to target node. transmission success ratio with different number of sen-
First, we simulate RCF, FAD, and DDST with dif- sor nodes. With the increase in sensor nodes, DDST
ferent communication radii and compare data has superiority on data transmission success ratio
Liang et al. 9

Network security analysis

Attackers aim to acquire unauthorized accessed data,
who are probably external intruders or network users.
Due to the lack of security in data transmission, nodes
in WSN are vulnerable to various illegal attacks by
deploying or capturing nodes. We assume illegally
deployed sensor nodes pretend to be normal network
nodes. These nodes could steal confidential data or per-
form false data injection attack. Besides, it is able for
attackers to capture many nodes in network. These
nodes can be utilized for sending plenty of false data to
consume network resource rapidly, such as energy,
bandwidth, computation ability, and storage. If the
transmitted data is attacked, we need to perform secu-
Figure 11. Comparison chart of transmission strategy on rity analysis on attacks like forging, signature tamper-
delivery cost. ing, and collusion.

(1) Resistance against forging attacks. First, data

transmitter sends a package including his false
name. Other nodes cannot forge this false
name; otherwise, it cannot pass the co-verifica-
tion. The difficulty to forge the identity is equal
to SHA-1 hash function.
(2) Resistance against signature tampering attacks.
The key for encryption is generated using
regeneration code. The privilege of decryption
is controlled. Even the data transmitter is
unable to tamper data. Besides, attackers can-
not get information about the key. So, the data
will not be forged and tampered. The key is
generated in authentication of multiple net-
work nodes. Single-node authentication cannot
tamper the signature.
Figure 12. Comparison chart of transmission strategy on
delivery delay. This article presents a distributed data
transmission approach to address above issues. Each
transmission of data package will generate a new path.
compared with other schemes because DDST utilizes Attackers can capture various data packages only by
analysis of relative movement distance between nodes capturing or eavesdropping all possible routing paths.
and transfer time competitive mechanism. The multiple paths are generated by all participated
The performance on delivery cost and delivery delay nodes. By capturing a part of nodes will not change the
is shown in Figures 11 and 12, which are better than result. So, it satisfies both the high efficiency and
other schemes. As shown in Figure 11, we assume the controllability.
energy is unlimited. Under the same condition, the pro- We conduct data capturing experiments on three
posed DDST protocol has the minimum cost compared schemes (RCF, FAD, and DDST) to verify the rela-
to other algorithms. Without the limitation of resource, tionship between probability of capturing data and net-
it is unnecessary to reduce overhead at the cost of trans- work size N in WSN. Figure 13 shows the experimental
mission success ratio or delay time. RCF and FAD result. Network size changes from 2 to 12. When other
have high energy consumption and delivery delay. The conditions are the same, various values of N have direc-
use of secure model of node competition reduces fre- tive impact on RCF and FAD. The effect on DDST is
quency of using sensor nodes. So, the delivery cost and lower. We have analyzed the reason. The larger N is,
delivery delay of DDST are the minimum. The compar- the more the created disjoint paths of DDST are. So,
ison of three indicators proves the transmission perfor- the security is higher. Although larger value of N can
mance of DDST is encouraging. improve performance of RCF and FAD, DDST has
10 International Journal of Distributed Sensor Networks

considered in secure data transmission. It reduces net-

work energy consumption and prolongs network life-
time. The nodes in WSN have complex features,
causing great difficulty in research. Compared to other
wireless networks, the study is still in its infancy. So,
research on secure data transmission in WSN will be a
big challenge in future works.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.

Funding
Figure 13. Relationship between probability for capturing data The author(s) disclosed receipt of the following financial sup-
and network size N. port for the research, authorship, and/or publication of this
article: This work was supported by the National Science
Foundation of China (grant no. 61572188), Hunan Provincial
better security because its path selection is controlled
Natural Science Foundation of China (grant no. 2016jj2058),
by original node and intermediate node.
Scientific Research Project of Hunan University of Science
and Technology (grant no. E51697), and Scientific Research
Fund of Hunan Provincial Education Department (grant no.
Conclusion 14A047).
This article targets at security in data transmission and
proposes a DDST approach. DDST divides data trans- References
mission into two stages, namely, constructing a collec- 1. Cui L, Ju HL, Miao Y, et al. Research on wireless sensor
tion of receiving nodes and selecting a unique network. J Comput Res Dev 2005; 42: 163–174.
forwarding node from this collection. DDST uses rela- 2. Xie K, Wang X, Wen JG, et al. Cooperative routing
tive movement analysis and competitive mechanism with relay assignment in multi-radio multihop wire-
based on forwarding time. This protocol achieves effec- less networks. IEEE/ACM T Networking 2016; 24:
tive, rapid, and precious data transmission. 859–872.
The main work is concluded as follows: (1) select 3. Mongiovi M, Singh AK, Yan X, et al. Efficient multi-
proper nodes as data receivers among all neighboring casting for delay tolerant networks using graph indexing.
nodes of the sender. The following two conditions are In: Proceedings of the 2012 international conference on
computer communications, 2012, pp.22–30. Santa Bar-
satisfied: candidate node will not move out the commu-
bara, CA: ACM Press, http://www-bcf.usc.edu/~kpsou-
nication range in data transmission period and candi-
nis/Papers/indexing.pdf
date node should be set within the intersection of 4. Zhang S, Cao J, Chen L, et al. Locating nodes in mobile
communication area and the 60-degree-sector-oriented sensor networks more accurately and faster. In: Proceed-
sink node. The candidates should satisfy both condi- ings of the 5th annual IEEE communications society con-
tions. (2) The competitive mechanism based on for- ference on sensor mesh and ad hoc communications and
warding time is used. After receiving data, the networks, San Francisco, CA, 16–20 June 2008, pp.37–
mechanism will compute forwarding competitive time 45. New York: IEEE.
itself with remaining energy Er and distance to sink 5. Zhang S, Cao J, Chen L, et al. Accurate and energy-
node. The candidate with the minimum competitive efficient range-free localization for mobile sensor net-
time will be selected to forward data. Other nodes will works. IEEE T Mobile Comput 2010; 9: 897–910.
6. Xing G and Wang T. Rendezvous planning in wireless
drop the received data.
sensor networks with mobile elements. IEEE T Mobile
At present, data transmission in network mainly
Comput 2008; 7: 1430–1443.
focuses on a single node. Data transmission among all 7. Li G, Shi W and Wang Y. Automatic ARIMA time
nodes has various energy, computation ability, storage, series modeling for data aggregation in wireless sensor
sensing areas, and communication distances. So, we networks. Int J Digital Content Tech Appl 2012; 6:
should simplify this issue to master the essence. The 438–447.
parameters of nodes in WSN are not always consistent 8. Feng Y, Gong H, Fan M, et al. A distance-aware replica
due to various factors. In this case, reasonable alloca- adaptive data gathering protocol for delay tolerant
tion and scheduling on work period of nodes should be mobile sensor networks. Sensors 2011; 11: 4104–4117.
Liang et al. 11

9. Lan F and Su W. A survey on sensor networks. IEEE 22. Heinzelma WR, Chandrakasan A and Balakrishnan H.
Commun Mag 2002; 40: 102–114. Energy-efficient communication protocol for wireless
10. Zhang S, Zeng Y, Chen L, et al. Performance evaluation microsensor networks. In: Proceedings of the 33rd Hawaii
of localization algorithms for mobile sensor networks. J international conference on system sciences, Maui, HI, 7
Software 2011; 22: 1597–1611. January 2000, p.223. New York: IEEE.
11. Guo LJ, Raheem A and Li YS. SMITE: a stochastic com- 23. Mohanty P and Kabat MR. Energy efficient reliable
pressive data collection protocol for mobile wireless sen- multi-path data transmission in WSN for healthcare
sor networks. In: Proceedings of the IEEE INFOCOM, application. Int J Wireless Inform Network 2016; 23:
Shanghai, China, 10–15 April 2011, pp.1611–1619. New 162–172.
York: IEEE. 24. Srouji MS, Bonny T and Henkel J. High-speed encoding/
12. Haneef M and Deng Z. Design challenges and compara- decoding technique for reliable data transmission in wire-
tive analysis of cluster based routing protocols used in less sensor networks. In: Proceedings of the 2014 11th
wireless sensor networks for improving network life time. annual IEEE international conference on sensing, commu-
Adv Inform Serv Sci 2012; 4: 450–459. nication, and networking, Singapore, 30 June–3 July 2014,
13. Zhang J and Li L. Maximum life-time broadcast authen- pp.1449–1454. New York: IEEE.
tication protocol for wireless sensor networks. Adv Inform 25. Marappan P and Rodrigues P. An energy efficient rout-
Serv Sci 2012; 4: 146–154. ing protocol for correlated data using CL-LEACH in
14. Katti S, Rahul H and Hu W. XORs in the air: practical WSN. Wirel Netw 2016; 22: 1415–1423.
wireless network coding. IEEE T Netw 2008; 16: 497–510. 26. Rani S, Malhotra J and Talwar R. Energy efficient chain
15. Wark T, Corke P, Sikka P, et al. Transforming agricul- based cooperative routing protocol for WSN. Appl Soft
ture through pervasive wireless sensor networks. IEEE Comput 2015; 35: 386–397.
Pervas Comput 2007; 6: 50–57. 27. Kim DS and Chung YJ. Self-organization routing
16. Li Z, Li M, Wang J, et al. Ubiquitous data collection for protocol supporting mobile node for wireless sensor net-
mobile users in wireless sensor networks. In: Proceedings work. In: Proceedings of the first international multi-
of the IEEE INFOCOM, Shanghai, China, 10–15 April symposiums on computer and computational sciences,
2011, pp.2246–2254. New York: IEEE. Hangzhou, China, 20–24 June 2006, pp.622–626. New
17. Wang Y, Wu H, Dang H, et al. Analytic, simulation, and York: IEEE.
empirical evaluation of delay/fault-tolerant mobile sensor 28. Wang Y and Wu H. DFT-MSN: the delay/fault tolerant
networks. IEEE T Wirel Commun 2007; 1: 3287–3296. mobile sensor network for pervasive information gather-
18. Maurice J, Chadi M and Fawaz F. Disruption-tolerant ing. IEEE T Mobile Comput 2007; 6: 1021–1034.
networking: a comprehensive survey on recent develop- 29. Li L, Jiang W, Sun L, et al. A receiver-based cross-layer
ments and persisting challenges. IEEE Commun Surv Tut forwarding protocol for mobile sensor networks. Chin J
2012; 14: 2–15. Comput Res Dev 2009; 46: 120–128.
19. Sha L and Gup T. An adaptive and efficient data delivery 30. Lu W and Li G. A prediction-based and low redundancy
scheme for DFT-MSNS (delay and disruption tolerant data forward strategy in mobile wireless sensor networks.
mobile sensor networks). In: Proceedings of the IEEE J Comput Res Dev 2011; 48: 125–130.
international conference on advances in engineering, science 31. Wang Y and Wu H. Replication-based efficient data
and management, Nagapattinam, India, 30–31 March delivery scheme (RED) for delay/fault tolerant mobile
2012, pp.99–104. New York: IEEE. sensor network (DFT-MSN). In: Proceedings of the 4th
20. Zhang K, Zeng J and Liu W. Data delivery scheme of annual IEEE international conference on pervasive comput-
DTMSN based on probability replication and its perfor-
ing and communications workshops, Pisa, 13–17 March
mance research. J Electron Inform Tech 2010; 32:
2006, pp.485–489. New York: IEEE.
677–681.
32. Blum B, He T, Son S, et al. A robust state-free communi-
21. Carlo C, Rosario F and Marco L. DTN bundle layer over
cation protocol for sensor networks. Charlottesville, VA:
TCP: retransmission algorithms in the presence of chan-
University of Virginia, 2003.
nel disruptions. J Commun 2010; 5: 106–116.