Malicious Node Detection for the Future Network Security from Epistemic Uncertainties

Xu Huang,Muhammad R. Ahmed,Hongyan Cui,Li Shutao
2013-01-01
Abstract:The next generation of WSN will benefit when sensor data is added to blogs, virtual communities, and social network applications. Wireless sensor networks (WSNs) are popular now as distributed networks exposed to an open environment, collecting of self-organized nodes with limited computation and communication capabilities and energy covering deployed areas that interested by the controllers. WSNs have been as aware of environmental technologies such as sound, water contamination, temperature, pressure, motion and other pollutants. However, as wireless communication becomes all sectors of daily life, the security threats to WSNs become increasingly diversified, prevention based due to the open nature of the wireless medium. The node in a WSN is called compromised becomes another major problem of WSN security because it allows an adversary to enter inside the security perimeter of the network and launch attacks, which has raised a serious challenge for WSNs. This paper is focusing on investigating integrating wireless sensor networks with cloud computing with the case that internal attacks of WSNs with multi-hop and single sinker for the future network, such as compromised nodes in a deployed WSN, where we first present our novel protecting algorithm to WSN to protect it with the evidences that our novel algorithm works efficiently and effectively. Our new algorithm takes advantages from uncertain decisions, which means there is no need to know the structure before we check the network. It is involved in the posteriori probability of binary events represented by the beta family of density functions and Dempester Shafer theory (DST). AS it has no need to have any knowledge about the structure of the network, it will be flexible to use in real life.
What problem does this paper attempt to address?