Summary: | 碩士 === 國立清華大學 === 資訊工程學系 === 96 === Wireless sensor networks are widely utilized in many areas nowadays. Sensor nodes have limited computation ability, memory and energy and so on. In addition, the energy consumption of communication is much more than computation for wireless sensor nodes. For the reason, data aggregation is a widely used technique in wireless sensor networks to reduce communication cost. Nevertheless, data aggregation brings some side effects and suffers from some attacks.
Firstly, data aggregation exploits fewer messages to represent a great deal sensing readings. Therefore, the aggregated values just present some results. Nevertheless, the base station is unable to be aware of the distribution of these original sensing values. In other words, the base station can not recover to original messages from aggregated value to perform further analysis.
Secondly, in wireless sensor networks, internal nodes are able to alter messages when it aggregates receiving messages. The altered messages will result in a deviation of final result. Consequently, the false result could cause a wrong decision. Hence, it is significant to protect these messages from an attacker altering them when internal nodes aggregate these data. However, less of them can indeed guarantee data integrity and resist dropping attack and so on. An attacker can simply forge data to affect the decision of base station by compromised nodes.
This thesis, therefore, proposes a lightweight secure data aggregation protocol to solve above side effects and attack models. In the meanwhile, this thesis also provides detailed security analyses to verify the security of our scheme. Finally, we compare our scheme with other precious methods in terms of security, memory cost, communication cost, computation cost, and application environment etc.
|