16-10-2012, 11:00 AM
BECAN: A Bandwidth-Efficient Cooperative Authentication Scheme for Filtering Injected False Data in Wireless Sensor Networks
BECAN.pdf (Size: 1.16 MB / Downloads: 53)
Abstract
Injecting false data attack is a well known serious threat to wireless sensor network, for which an adversary reports bogus
information to sink causing error decision at upper level and energy waste in en-route nodes. In this paper, we propose a novel
bandwidth-efficient cooperative authentication (BECAN) scheme for filtering injected false data. Based on the random graph
characteristics of sensor node deployment and the cooperative bit-compressed authentication technique, the proposed BECAN
scheme can save energy by early detecting and filtering the majority of injected false data with minor extra overheads at the en-route
nodes. In addition, only a very small fraction of injected false data needs to be checked by the sink, which thus largely reduces the
burden of the sink. Both theoretical and simulation results are given to demonstrate the effectiveness of the proposed scheme in terms
of high filtering probability and energy saving.
INTRODUCTION
DUE to the fast booming of microelectro mechanical
systems, wireless sensor networking has been subject to
extensive research efforts in recent years. It has been well
recognized as a ubiquitous and general approach for some
emerging applications, such as environmental and habitat
monitoring, surveillance and tracking for military [1], [2], [3],
[4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16]. A
wireless sensor network is usually composed of a large
number of sensor nodes which are interconnected through
wireless links to perform distributed sensing tasks. Each
sensor node is low-cost but equipped with necessary sensing,
data processing, and communicating components. Therefore,
when a sensor node generates a report after being
triggered by a special event, e.g., a surrounding temperature
change, it will send the report to a data collection unit (also
known as sink) through an established routing path [17]
MODEL AND DESIGN GOAL
In this section, we formulate the network model, the
security model, and identify the design goal.
2.1 Network Model
Weconsider a typical wireless sensor network which consists
of a sink and a large number of sensor nodes N ¼
fN0;N1; . . .g randomly deployed at a certain interest region
(CIR) with the area S. The sink is a trustable and powerful
data collection device, which has sufficient computation and
storage capabilities and is responsible for initializing the
sensor nodes and collecting the data sensed by these nodes.
Each sensor node Ni 2 N is stationary in a location. For
differentiation purpose, we assume each sensor node has a
unique nonzero identifer. The communication is bidirectional,
i.e., two sensor nodes within their wireless transmission
range ® may communicate with each other. Therefore,
if a sensor node is close to the sink, it can directly contact the
sink. However, if a sensor node is far from the transmission
range of the sink, it should resort to other nodes to establish a
route and then communicate with the sink. Formally, such a
wireless sensor network.
Security Model
Since a wireless sensor network is unattended, a malicious
adversary may readily launch some security attacks to
degrade the network functionalities. In addition, due to the
low-cost constraints, sensor nodesN ¼fN0;N1; . . .g are not
equipped with expensive tamper-proof device and could be
easily compromised in such an unprotected wireless sensor
network. Therefore, in our security model, we assume an
adversary A can compromise a fraction of sensor nodes and
obtain their stored keying materials. Then, after being
controlled and reprogrammed by the adversary A, these
compromised sensor nodes can collude to launch some
injected false data attacks.
Since our work focuses on filtering injected false data
attack, other attacks launched by the compromised sensor
nodes in wireless sensor network, such as building bogus
routing information, selectively dropping true data packet,
and creating routing loops to waste the energy of network
[18], are not addressed in this paper.
RELATED WORK
Recently, some research works on bandwidth-efficient
filtering of injected false data in wireless sensor networks
have been appeared in the literature in [9], [10], [11], [12],
[13]. In [9], Ye et al. propose a statistical en-routing filtering
mechanism called SEF. SEF requires that each sensing report
be validated by multiple keyed message authenticated
(MACs), each generated by a node that detectes the same
event. As the report being forwarded, each node along the
way verifies the correctness of the MACs at earliest point. If
the injected false data escapes the en-routing filtering and is
delivered to the sink, the sink will further verify the
correctness of each MAC carried in each report and reject
false ones. In SEF, to verify the MACs, each node gets a
random subset of the keys of size k from the global key pool
of size N and uses them to producing the MACs. To save the
bandwidth, SEF adopts the bloom filter to reduce the MAC
size. By simulation, SEF can prevent the injecting false data
attack with 80-90 percent probability within 10 hops.
However, since n should not be large enough as described
above, the filtering probability at each en-routing node p ¼
kðTNcÞ
N is relatively low. Besides, SEF does not consider the
possibility of en-routing nodes’ compromise, which is also
crucial to the false data filtering. In [10], Zhu et al. present an
interleaved hop-by-hop authentication (IHA) scheme for
filtering of injected false data. In IHA, each node is associated
with two other nodes along the path, one is the lower
association node, and the other is the upper association node.
An en-routing node will forward received report if it is
successfully verified by its lower association node. To reduce
the size of the report, the scheme compresses t þ 1 individual
MACs by XORing them to one. By analyses, only if less than t
nodes are compromised, the sink can detect the injected false
data. However, the security of the scheme is mainly
contingent upon the creation of associations in the association
discovery phase.
CONCLUSION
In this paper, we have proposed a novel BECAN scheme for
filtering the injected false data. By theoretical analysis and
simulation evaluation, the BECAN scheme has been
demonstrated to achieve not only high en-routing filtering
probability but also high reliability with multi-reports. Due
to the simplicity and effectiveness, the BECAN scheme
could be applied to other fast and distributed authentication
scenarios, e.g., the efficient authentication in wireless
mesh network [31]. In our future work, we will investigate
how to prevent/mitigate the gang injecting false data attack
from mobile compromised sensor nodes [32].