Abstract

Abstract: Due to the sensitive nature of the data gathered by many wireless sensor networks (WSNs) it is becoming critical that this data be protected. However, due to the constrained nature of resources on sensor nodes, this is a difficult task. In particular, the use of asymmetric cryptographic operations, i.e. public key ciphers, often places an unjustifiable burden on a sensor node’s resources. As a result symmetric key ciphers are primarily used in WSNs. This introduces the difficult task of deploying and managing the required symmetric keys, which can be a major challenge even for moderately sized networks. All currently available WSN specific solutions to this problem either have a very simple key utilisation strategy for the network, resulting in a low level of security overall, or else only provide limited connectivity. Additionally the majority of these solutions are overly complex, both conceptually and in terms of implementation, and so they are not used. This work identifies ten requirements for a WSN key management solution and then presents the design, implementation and evaluation of a solution, called µKM, which meets each of these requirements and overcomes the problems of the existing schemes. This is achieved by relaxing the memory constraint in order to provide a large pool of keys to each node, a valid concession on newer generation sensor nodes. The evaluation of µKM shows that it is as efficient, if not more so, than the existing solutions in terms of energy consumption, network latency, and, to a lesser extent, program memory and RAM requirements. It also comes out well ahead of the alternatives in link key establishment overheads due to the fact that it requires no prior and/or additional communication in order to set up individual link keys between any two nodes.

Bio: Dr. Michael Healy is the lead embedded systems software developer for Shimmer Research, a supplier of wireless sensor network technology primarily focused on health and fitness applications. He received a BEng Degree in Computer Engineering from the University of Limerick in 2005 and was granted a PhD from the same institution in 2012 for work on securing wireless sensor networks. Prior to Shimmer Research Michael worked as an R&D applications engineer in Intel’s digital health group.