Research On Random Key Pre-distribution And Anonymity Technology In Wireless Sensor Networks

Wireless sensor networks(WSNs)are easy to deploy and can collect surrounding data conveniently.With rapid development of wireless communication and sensor technology,wireless sensor networks have been widely used in environmental monitoring,military,industrial control fields.Sensor nodes can cooperate to transmit sensed information(e.g.,temperature,humidity)to a base station.As sensor nodes' energy is limited,if they are deployed in harsh or hostile environment,it is difficult to replace nodes' batteries.In addition,sensor nodes are vulnerable to hostile attacks,yielding identity forgery,violation of data privacy and other security issues.To protect security of wireless sensor networks,it is necessary to provide suitable key management schemes for WSNs.Traditional public key architecture is not applicable for WSNs since overhead of public key operations is heavy for sensor nodes.This dissertation focus on design of WSN random key pre-distribution scheme based on cluster structure,anonymity for key update process and source nodes' location protection.The main contribution of this dissertation is summarized as follows:1.EG random key pre-distribution scheme is designed for flat structure and its node storage overhead increases with network size.An asymmetric random key pre-distribution scheme based on clustering structure is proposed in which complete key sharing is adopted for cluster head nodes while asymmetric key ring size is assigned for cluster head nodes and its member nodes.This asymmetric structure can improve the probability of key sharing among cluster head nodes and its member nodes to establish intra-cluster secure connections.As cluster head nodes are guaranteed to be securely connected,inter-cluster secure connections are established by relay through cluster head nodes.Analysis shows that compared with EG scheme,this asymmetric scheme can effectively reduce key storage cost of nodes while maintaining the same safe connection probability between nodes.2.EG random key pre-distribution scheme assumes nodes exchange key identifiers to determine common keys between them after assigning key rings for nodes.However,an adversary can learn topology information of the underlying random key graph by intercepting exchanged key identifiers.In addition,when key exposure occurs,compromised nodes should be revoked and uncompromised nodes' key rings should be updated securely.An anonymous key update mechanism based on bloom filter is presented that can revoke compromised nodes while an adversary is infeasible to link key identifiers with a node.3.Sensor node transmitting data is vulnerable to localization and hostile attack.A Source node location protection scheme based on backbone structure is proposed in this dissertation,that is,an attacker cannot determine whether a node is reporting sensed events.At first,a dominating set is constructed as backbone access structure by running node coloring algorithm.A source node routes its data packet to the backbone access structure by random walk.In order to hide source node's information who is reporting sensed event,other nodes generate virtual events with a certain probability when there is no sensed event to report in each round.Events are encrypted and relayed to nearby backbone nodes.The corresponding backbone nodes randomly select proxy backbone nodes to forward data packets to sink,making it difficult for an attacker to correctly locate source nodes.