IJRET: International Journal of Research in Engineering and Technology

eISSN: 2319-1163 | pISSN: 2321-7308

SECURE DATA DISSEMINATION PROTOCOL IN WIRELESS SENSOR

NETWORKS USING XOR NETWORK CODING
Jisha Mary Jose1, Jomina John2
1
2

Dept. of CSE, Rajagiri School of Engineering and Technology, Kochi, India

Dept. of CSE, Rajagiri School of Engineering and Technology, Kochi, India

Abstract
Wireless sensor networks (WSN) are basically distributed networks or a collection of sensor nodes which collect information
which are used to analyse physical or environmental conditions. WSNs are usually setup in remote and hostile areas and work in
extreme conditions. Applications of WSN include habitat monitoring, industrial applications, battlefield surveillance, smart homes
etc. Most of them require regular updating of software in sensor nodes through the wireless channel for efficient management and
working. So it is necessary to spread data through the wireless medium after the nodes are deployed. This is known as data
dissemination or network reprogramming. A good data dissemination protocol must be fast, secure, reliable and energy efficient.
To achieve these we can make use of network coding techniques which reduces the number of retransmissions due to any packet
drops. But network coding increases the chance of various kinds of network attacks. Also to avoid spreading of malicious code in
the network, each sensor node has to authenticate its received code before propagating it further. So here a novel dissemination
protocol is introduced based on simple cryptographic techniques which prevents pollution and DoS attacks and at the same time
achieves fastness using the technique of network coding.

Keywords: wireless sensor network; dissemination; reprogramming; network coding; security; pollution attack.
--------------------------------------------------------------------***---------------------------------------------------------------------1. INTRODUCTION
Wireless Sensor Networks (WSN) is one of the major
milestones in the field of communication. These networked
collection of nodes take us a step closer to obtaining
valuable information about the physical world. WSN are
used popularly in many applications like remote control and
monitoring, construction safety systems, environmental
monitoring, health care management, disaster management,
surveillance operations, smart homes, habitat monitoring,
indoor sensor networks, seismic monitoring of buildings etc
[1]. In computer science and communication wireless sensor
networks entertain lot of research today.
A WSN is made of sensor nodes used for monitoring and
analysis purposes as shown in Fig 1. These sensor nodes
pass the information that they collect to a prime location
called a base station. In most systems, a WSN
communicates with a LAN or WAN through a gateway like
medium. The gateway is actually a bridge between the WSN
and the various other networks [2]. This allows data to be
stored by devices and which can be taken up for processing
later. Each sensor node or mote has several parts: a circuit
for interfacing with other sensor nodes, a micro controller, a
radio transceiver, and a battery for power supply. The
topology used can be either a star, ring, grid network or
multi-hop wireless mesh network.WSN is used primarily in
remote and hostile environments for information collection.
Hence it is a major challenge to produce cheap sensor nodes.
They must be designed carefully by considering all the
different constraints of the environment in consideration.

Fig -1: An example wireless sensor network

Wireless sensor networks must be operated for long duration
of time and usually dont get any human administration or
intervention in between [3]. Evolving conditions and
environments can also; cause changes in application
features, which hence lead to the need to change the network
behaviour by introducing new code or software. But the
remote nature of WSN is a disadvantage here. It will require
the propagation of new code updates over the wireless
medium i.e. over the air as manual updating of such
networks will not be possible. This process is known as
dissemination or network reprogramming.
But dissemination brings in a lot of challenges [3]. One
major challenge is proper and complete dissemination of
information to all sensor nodes in the network. This is

_______________________________________________________________________________________
Volume: 03 Special Issue: 15 | Dec-2014 | IWCPS-2014, Available @ http://www.ijret.org

57

IJRET: International Journal of Research in Engineering and Technology

difficult since the number of nodes in the network can be

huge and the environment is always dynamic, thus the basic
topology keeps changing constantly. Secondly the
information to be disseminated may be produced at a single
node, such as the prime source i.e. the base station, or at the
sensor nodes themselves. Thirdly data must be disseminated
in a secure way or else adversaries can track out critical
data. Also there is a possibility of attackers sending bogus
data into the network which must not be received by the
sensor nodes as they can cause different attacks like
pollution attacks, denial of service attacks and so on. So
dissemination of code or program data in wireless sensor
networks is an area to be worked in deeply and new
techniques need to be introduced to achieve tradeoffs
between energy and speed in dissemination. The aim of this
work is to develop a novel secure and fast data
dissemination protocol for use in wireless sensor networks.
This work concentrates on developing a dissemination
protocol for dissemination of small data.
Linear network coding is a technique used to achieve
fastness and energy efficiency during dissemination [4]. It is
a technique that combines packets in the network; increasing
the throughput, decreasing energy consumption, and
reducing the number of messages transmitted. In traditional
systems dropped packets are recovered using
retransmissions. But in network coding we can combine
packets using mathematical operations and then disseminate
so that recovery of lost packets can be achieved without
retransmission.
But network coding along with its energy efficiency
advantages brings in a lot of headaches. It is highly prone to
attacks like pollution, denial-of-service attacks and many
others. So to deal with these the proposed system uses
simple but efficient cryptographic techniques for data
dissemination. This makes sure that we can achieve simple
yet secure data dissemination in wireless sensor networks.
Our work is organized as follows. First we focus on the need
of data dissemination in wireless sensor networks and some
of its related works. Next the design and implementation
which concentrates on the dissemination of small values and
variables is explained. Then we study the performance of
the new protocol through extensive simulation using
TinyOS and finally have the conclusion and references.

2. RELATED WORKS
Data dissemination in wireless sensor networks is a critical
and vital task. It is based on the concept of traditional
communication system, where we have a sender and
receiver. The scenario is basically a sender sending out
some information, and receiver collecting the information
sent, processing it and sending some information back.
While in data dissemination, only half of this concept is
applied. Some information is sent out and received at the
destination, but no reply is given back. The sender sends out
information, not to one node, but to many as in a
broadcasting system.

eISSN: 2319-1163 | pISSN: 2321-7308

Dissemination is used to send code updates or program

images to the sensor nodes periodically so as to perform
reprogramming of the nodes. This over the air act is required
since manual updating of sensor nodes deployed in remote
environments is next to impossible in most of the cases. The
main aim of a dissemination protocol in WSN is to ensure
that all the sensor nodes have consistent data with them
always.
There are two kinds of dissemination in WSN [5]:

Code dissemination - to send program images which

are generally bulky data. Usually they are divided
into fixed sized pages and packets and then
disseminated.

Data discovery and dissemination - to disseminate

small configuration parameters, variables, queries,
commands etc in packets.

Fig -2: Dissemination process in WSN

2.1 Small Value Dissemination

This work concentrates on data discovery and dissemination
protocols i.e. dissemination of small values like variables,
parameters and so on. Figure 2 gives a general idea about
data dissemination. Traditional protocols available for this
include Drip, DIP and DHV. They are all based on Trickle
algorithm [6].
Drip proposed by Tolle et. al [7] is the simplest of all
dissemination protocols and is based on Trickle algorithm
and establishes an independent trickle for each variable in
the data. Every time an application wants to transmit a
message, a new version number is generated and used. This
will cause the protocol to reset the Trickle timer and thus
disseminate the new value else the trickle timer interval is
incremented.
DIP (DIssemination Protocol) [8] is a data detection and
dissemination protocol proposed by Lin et al. It is a protocol
based on the Trickle algorithm. It works in two parts:
determining whether there a difference in data stored at a
node, and then determining which data is different. It is
based on the concept of version number and key tuple for
each data item. DIP calculates hashes that cover all version
numbers of the data. Nodes that receive hashes same as their
own know that they have consistent data with respect to
their nearest neighbours. If a node gets a hash that differs
from its own hash, it knows that a difference exists in the
data.

_______________________________________________________________________________________
Volume: 03 Special Issue: 15 | Dec-2014 | IWCPS-2014, Available @ http://www.ijret.org

58

IJRET: International Journal of Research in Engineering and Technology

DHV (Difference detection-Horizontal search-Vertical

search), [9] is a code consistency maintenance protocol
given by Dang et al. It tries to keep codes on different nodes
in a WSN consistent and up to date. Here also data items are
represented as tuples (key, version). It is based on the
observation that if two data items are different, they will
only differ in a few least significant bits (LSB) of their
version number rather than in all their bits. So only those
bits need to be checked. For this steps followed are detection
and identification.
There are many code dissemination protocols like Deluge
[10] as well. They are used to disseminate large code
updates into the network. For this the code is often broken
down into pages and then further into packets. Here we
have seen some basic data discovery and dissemination
protocols. They dont support any techniques which help to
reduce packet retransmissions. Also none of these protocols
provide security to the data disseminated.

2.2 Network Coding and Data Dissemination

Network coding [4] aims to replace the traditional store and
forward technique used in networks; by better routing
algorithms that will allow intermediate nodes to transform
the moving data. Network coding has become popular due
to its properties like robustness and better throughput. It
helps to achieve fast data dissemination as it reduces the
number of retransmissions that will be needed if there are
packet losses. Many dissemination protocols have been
developed using the concept of network coding [14] [15].
The advantages of network coding based dissemination
protocols are that they achieve energy savings and
communication efficiency, especially during increased
packet loss or network density. So network coding based
protocols can be profitable for reprogramming of WSNs.
However we face a potential problem in hostile
environments. An adversary may launch pollution attacks, in
which a malicious node sends bad encoded packets that
consist of bogus data, which leads to incorrect decoding of
the original data upon retrieval.
Here we use binary network coding i.e. the mathematical
operation used is XOR to combine the contents of packets
[4]. Only two packet network coding is done here. Also here
we focus on dissemination of small values like configuration
parameters, variables, queries, commands etc whose size
range from 2-4 bytes and thus is modification of the existing
DRIP protocol.

3. ASSUMPTIONS AND THREAT MODEL

3.1 Assumptions
Here assume that the source of the reprogramming variables,
i.e., the base station, is a secure location. Also each sensor
node has a unique identification number. We assume that
while each sensor node is resource limited, it has sufficient
memory to store all the security mechanisms of the protocol.

eISSN: 2319-1163 | pISSN: 2321-7308

3.2 Threat Model

Here assume that the individual sensor-nodes are
unprotected. An adversary may insert its own attacker nodes
into the network, or it may capture other nodes. The
adversary can attempt to launch pollution attacks to corrupt
the data in the network and also to consume the limited
resources on sensor nodes.

4. PROPOSED SYSTEM
In this protocol data dissemination is done in a secure and
fast way by using the techniques of network coding and
cryptography. Network coding reduces the number of
retransmissions due to any packet losses happening in the
network by combing and sending data. Also data
disseminated is always sent as encrypted data. For this
nodes first perform node to node authentication and
establish session keys. Then the session key is used for
encrypted transfer of data. This protocol ensures that the
system is free of pollution [13] and Denial-of-Service
attacks. The different phases of this protocol include:

4.1 System Initialization Phase

This phase is done before the WSN is deployed in the
application field. In this phase the base station generates a
master key Km and a unique random number Rm and stores
them safely in each node. Also a list of all the valid node ids
is maintained in each node.

4.2 Packet Processing Phase

In this phase the actual data dissemination occurs. Before
disseminating data a node will generate a real time key
using a key generation algorithm. This includes the
generation of two unique random numbers R1_node and
R2_node. Key generation is done using Trivium-Multilinear
Modular Hashing (MMH) as the MAC function and SHA1
as hashing function H(x). The steps are:
1.

MAC[i] = R1_node XOR K[i]

(1)

2.

a[i] = node_id + MAC[i]

(2)

3.

h = MMH ( a[i] )

(3)

4.

Key = H ( h XOR R2_node )

(4)

Where K[i] is the master key of the MAC function, node_id

is the identifier of the corresponding node, XOR is the
logical XOR operation. This real time key is broadcast by
the node in a packet which will include the node_id and the
key. The destination node who receives it will check the
node_id with its list of valid nodes and ensure this packet is
coming from a valid node. If yes that node will also generate
a real time key using the same process as above and send
back a reply packet to the sender node which will contain
the node_id and the newly generated key.

_______________________________________________________________________________________
Volume: 03 Special Issue: 15 | Dec-2014 | IWCPS-2014, Available @ http://www.ijret.org

59

IJRET: International Journal of Research in Engineering and Technology

If this packet is also validated, then the two nodes are ready
to generate a session key. The key is generated as:
Session key = Km XOR Ka XOR Kb

(5)

Where Ka and Kb are keys generated at any two nodes A and

B. Now this key is used for encrypting the data to be
disseminated. The advantage of this scheme is that there is
no need of actual exchange of the session key through the
network. To encrypt the data we use symmetric encryption
techniques preferably Advanced Encryption Standard
(AES). So the data packet disseminated from a node will
contain the data in encrypted form i.e.
Data = E (d)sk , where sk is the session key.

(6)

Dissemination in wireless sensor networks works on the

basis of Trickle algorithm [2]. It takes on the concept of
gossiping. Whenever a new data is to be disseminated the
trickle timer is reset to 0 and the data is broadcasted. When a
node receives a new data it will store it. But if it receives a
data which it already is aware of then it will increase the
trickle timer interval and suppresses the duplicate incoming
data.
To achieve immediate authentication of data packets a
onetime hash of the initially generated random number is
also calculated and included in each packet. The steps are:
1.

Calculate Hash = H(Rm)

(7)

2.

Result = ADD(Hash)

(8)

Where H( ) is SHA-1 and ADD( ) is basic addition

operation. The result is included in the packets sent.

4.3 Packet Verification Phase

To achieve immediate authentication of the received packet,
the destination node will calculate the hash of Rm stored in
its memory and compare it with the value in the received
packet. If they match, then the received packet is a valid
node. Thus it will be acknowledged ACK by the destination.
Otherwise a NACK (negative ack) is sent to the sender.
Next we will have to ensure the integrity of the data. For this
first the node checks the id in the received packet. If it is a
valid node_id, then it will attempt to decrypt the data using
the session key already generated and stored. Every node
has an original data and combined data buffer. So the node
will check whether it is an original data or combined data. If
it is an original data it will be stored and disseminated after
a trickle timer fire and if it is a combined data, the node will
check whether it is possible to extract any other data from
this newly received data using network coding. After that
the data will be stored or disseminated out.

eISSN: 2319-1163 | pISSN: 2321-7308

makes sure that only valid data is sent out and data is been
sent out safely.

5. IMPLEMENTATION AND RESULTS

This protocol has been implemented in TinyOS-2.1.2
simulator TOSSIM [17]. We have considered a network
topology consisting of 100 nodes and 25 different data
variables are disseminated. The packet size in TinyOS [18]
is 29bytes. The sensor node considered for simulation here
is micaz.
Cryptographic support has been achieved using hashing
algorithms like SHA-1 which generates a 160 bit hash value,
MAC functions like Trivium Multilinear Modular Hashing
(MMH), and symmetric encryption algorithms like AES
which uses a 128 bit key.
The new protocol is found to resist cases of pollution attacks
i.e. only valid data packets are received and processed by
the intermediate nodes in the network. Also immediate
authentication of packets is achieved using the one time
hash value generated and stored in the data packets
disseminated.

6. SECURITY AND PERFORMANCE ANALYSIS

First we perform and analyze the security offered by this
protocol.

Resistance to pollution attacks- Attackers cant

pollute the network with bogus data since data
transfer done is always verified using cryptographic
techniques.

Resistance to Denial-of-Service attacks- Immediate

authentication of packets is done at each destination,
so bogus packets can be discarded and only valid
packets pass through.

Session key agreement- Session keys are used for

encryption and decryption. Also this key is locally
generated and used, hence not exchanged in the
network.

Real time key generation- No-pre stored keys in

nodes; they are calculated at time of data transfer
only.

Light-weight- Only simple yet good mathematical

operations and encryptions techniques are used
hence no much resource usage in nodes.
Chart 1 gives a comparison graph on the number of data
messages disseminated in each protocol namely DRIP [7],
CodeDrip [11] and the new proposed protocol. Network
coding has reduced total number of messages.

So likewise all the data disseminated from the original

source node will be distributed to all the nodes and a round
of dissemination will be completed. This technique thus

_______________________________________________________________________________________
Volume: 03 Special Issue: 15 | Dec-2014 | IWCPS-2014, Available @ http://www.ijret.org

60

IJRET: International Journal of Research in Engineering and Technology

[7]

[8]

[9]

[10]
Chart 1: Comparison of data messages disseminated

7. CONCLUSION
This paper proposes a novel data discovery and
dissemination protocol for wireless sensor networks which
can be used to achieve secure and fast data dissemination
especially for small configuration parameters and variables.
This technique combines the concepts of network coding
and simple cryptographic techniques so as to disseminate
data. The advantages of this protocol are that it is resistant to
pollution attacks, and achieves immediate authentication of
data been disseminated. Session keys are used to encrypt
and send data between nodes and there is no need of actual
transfer of the session keys through the network. Also only
simple mathematical operations are used to calculate keys
for encryption of data so not much of resource usage at the
nodes. All together it aims to provide a simple yet secure
and fast data dissemination protocol for usage in wireless
sensor networks. Node compromise by an attacker can be an
issue in this protocol. It will be dealt with as part of the
future works.

REFERENCES
[1]

[2]
[3]

[4]
[5]

[6]

Mohammad A. Matin, Wireless Sensor Networks:

Technology and Protocols: Published by InTech,
Croatia, ISBN 978-953-51-0735-4, 2012.
Salvatore La Malfa, Wireless Sensor Networks,
2010.
Jisha Mary Jose, Jomina John, Data dissemination
protocols in wireless sensor networks-a survey,
IJARCCE, March 2014.
T. Ho and D. Lun. Network Coding: An
Introduction. Cambridge University Press, 2008.
Daojing He, Sammy Chan, Shaohua Tang and
Mohsen Guizani, Secure Data Discovery and
Dissemination based on Hash Tree for Wireless
Sensor Networks, IEEE transactions on wireless
communications, Vol. 12, No. 9, September 2013.
P. Levis, N. Patel, D. Culler and S. Shenker,
Trickle: a self regulating algorithm for code
maintenance and propagation in wireless sensor
networks, in Proc. 2004 NSDI, pp. 15-28.

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]
[19]

eISSN: 2319-1163 | pISSN: 2321-7308

G. Tolle and D. Culler, Design of an applicationcooperative management system for wireless sensor
networks, in Proc. EWSN, pp. 121132, 2005.
Lin, K., Levis, P.: Data discovery and
dissemination with dip. In: Proceedings of the 2008
International Conference on Information Processing
in Sensor Networks (IPSN 2008), Washington, DC,
USA, IEEE Computer Society (2008) 433-444.
T. Dang, N. Bulusu, W. Feng, and S. Park, DHV: a
code consistency maintenance protocol for multihop wireless sensor networks, in Proc. 2009
EWSN, pp. 327-342.
Hui, J.W., Culler, D.: The dynamic behaviour of a
data
dissemination
protocol
for
network
programming at scale. In: Proceedings of the 2nd
international conference on Embedded networked
sensor systems (Sensys 04), New York, NY, USA,
ACM (2004) 81-94.
Nildo dos Santos Ribeiro Junior, Marcos A. M.
Vieira1, Luiz F. M. Vieira1 and Om Gnawali,
CodeDrip: Data Dissemination Protocol with
Network Coding for Wireless Sensor Networks, In
Proceedings of the 11th European conference on
Wireless sensor networks (EWSN 2014), Feb. 2014.
Hailun
Tan,
Secure
multi-hop
network
programming with multiple one-way key chains,
In: Proceedings of the International conference on
Embedded networked sensor systems (Sensys 07),
Sydney, Australia, ACM.
Yingpei Zeng, Jiannong Cao, Shigeng Zhang,
Shanqing Guo, Li Xie, Pollution Attack: A New
Attack Against Localization in Wireless Sensor
Networks, IEEE, WCNC-2009.
I-Hong Hou, Yu-En Tsai, T.F. Abdelzaher, and I.
Gupta. Adapcode: Adaptive network coding for
code updates in wireless sensor networks. In
INFOCOM 2008.The 27th Conference on Computer
Communications. IEEE, pages 15171525, 2008.
Andrew Hagedorn, David Starobinski, and Ari
Trachtenberg. Rateless deluge: Over-the-air
programming of wireless sensor networks using
random linear codes. In Proceedings of the 7th
international conference on Information processing
in sensor networks, IPSN 08, pages 457466,
Washington, DC, USA, 2008. IEEE Computer
Society.
S. Katti, H. Rahul,W. Hu, D. Katabi, M.Medard, and
J. Crowcroft, XORs in the air: practical wireless
network coding", ACM SIGCOMM Computer
Communication Review, vol. 36, no. 4, pp. 243-25,
2006.
Philip Levis, Nelson Lee, Matt Welsh, and David
Culler. Tossim: accurate and scalable simulation of
entire tinyos applications, In SenSys '03, pages 126137, New York, NY, USA, 2003. ACM Press.
TinyOS: an open-source OS for the networked
sensor regime. Available: http://www.tinyos.net/.
Simulating a Wireless Sensor Network, 2010-2013,
[Online] Available: http://virtual-labs. ac.in/ cse28/
ant/ ant/8/theory/.

_______________________________________________________________________________________
Volume: 03 Special Issue: 15 | Dec-2014 | IWCPS-2014, Available @ http://www.ijret.org

61