Security and Privacy in Body Sensor Networks - E-Healthcare Systems and Wireless Communications - page 124

signature structures. This is crucial in WSNs to
reduce the communication overhead and hence
the energy requirements of the sensor application.
TinyECC is configurable and optimized for
a wide range of sensor platforms. This can be
achieved explicitly by employing a set of software
optimization switches that can be controlled to
provide the best performance/resource utiliza-
tion combination. Currently TinyECC version
1.0 supports digital signatures using the ECDSA
algorithm, the ECDH key agreement protocol, and
the ECIES (Shoup, 2001) public key encryption
algorithm. For portability reasons, the library is
implemented in nesC (Gay et al., 2003), the Tin-
yOS default application programming language.
Moreover an inline assembly implementation is
provided for performance critical functions such as
big integer multiplications. The TinyECC library
can be downloaded from http://discovery.csc.ncsu.
edu/software/TinyECC/. The site contains detailed
description of TinyECC installation, performance
benchmarks, and optimization switches.
It is worth mentioning here that despite all
the performance optimizations provided by the
efficient ECC implementations in TinyECC, many
researchers believe that securing sensor applica-
tions using public-key cryptographic techniques
is still very computationally-intensive and thus
infeasible in embedded environments.

nodes and the specificity of the security protocol
requirements. The design guidelines are divided
into three main protocol categories:
Authentication and Key Management Proto-
cols: Perhaps the easiest and most efficient ap-
proach for key distribution in BSNs is to prepro-
gram the ciphering keys in the persistent
memory storage of the sensor nodes before de-
ployment in the human body. This is referred to
as the static key pre-distribution method discussed
in the fourth and ninth sections. The main problem
with this scheme is represented in the security
risks associated with the use of the same network-
wide keying material in the body sensor nodes.
Any attacker capable of capturing a node and
extracting the embedded cryptographic keys
would be able to execute different forms of inter-
ception, modification, and fabrication attacks
against the confidentiality, integrity, and authen-
ticity of communicated messages. If node capture
attacks are not feasible in the BSN operation
environment, then static key pre-distribution
would be the recommended method for executing
efficient key management and distribution in
BSNs.
Another efficient key agreement method relies
on the analysis of physiological body signals ex-
tracted from the human body such as ECG, inter-
pulse interval, and heart rate to derive the ciphering
keys of the sensor nodes. This is termed as the
human identification key agreement discussed in
the fourth section. Key agreement based on hu-
man identification techniques is highly suitable
in resource-limited body sensor networks with
low energy capabilities. However, this method
imposes a set of strict requirements on the sensor
nodes as well as on the properties of the extracted
body signal which renders it practically infeasible
in real BSN environments. Firstly, human iden-
tification key agreement requires that the sensor
nodes participating in the protocol be capable of
extracting the same physiological body signal or set
of signals with high degrees of accuracy. Secondly,
this method requires the presence of a time syn-
chronization protocol that facilitates the extraction

BLUEPRINT GUIDELINES
FOR DESIGNING BSN
SECURITY FRAMEWORKS

This section presents a summary of the main
security mechanisms and protocols described
earlier in this chapter. The presented model, shown
in the block diagram in Figure 8, represents a
generalized set of design patterns to be followed
when devising security protocols for protecting
BSN healthcare applications. Moreover it serves
as an adaptive blueprint that lays the ground for
the secure incorporation of protocol building
blocks based on the capabilities of the BSN sensor