Abstract:
In a battlefield surveillance scenario, data readings
and events emerging from a wireless sensor network deployed
that may not be used immediately by or simply impossible to
transmit to an authorized user (a Soldier) in real time are stored
in the network. Without proper protection for the sensitive data
generated in this setting, a compromised storage node (by an
enemy soldier) may divulge its stored sensitive data about the
monitored environment, and even worse, it may alter the data. In
this paper, we integrate an elliptic curve cryptography scheme
and an erasure coding scheme to provide reliable and secure
distributed in-network data storage for sensor networks. The
main idea is to distribute each erasure coded fragment appended
with a fingerprint to different storage nodes. The fingerprint is
to allow each coded data fragment to be independently verified
as a valid and correct subset of a specific data item. So, the
scheme achieves localization of data error. The proposed scheme
is resilient to collusion attack, pollution attack, and data dropping
attack, and guarantees forward and backward data secrecy
as well. The security of the proposed scheme relies on the
intractability of the elliptic curve discrete logarithm problem.
Different from the existing solutions, the uniqueness of our
method comes from the use of lightweight encryption scheme,
which is well suited for resource constrained wireless sensors.
