Research and Implementation on TinySec Rekeying for Wireless Sensor Networks
Tieming Chen, Dujuan Yan · 2008
Two neural network-based models can synchronize its initially different weight vectors by limited steps of single output bit-based mutual learning if only the two models hold a common input space. Such novel weight synchronization property can well be used for secure key exchange over a public communication channel. Itpsilas more suitable for the resource-constraint sensor networks because that the full weight synchronization model can be easily implemented with low consuming of both computing and communication. At current, as for wireless sensor networks, TinySec is a popular deployed protocol for data link security. However, there is no key management solution designed for TinySec protocol, such as shared key distribution and update. In this paper, a lightweight key update scheme for TinySec is proposed based on the weight synchronization model. The implementations, as well as simulation tests, for the proposed scheme are all discussed.