Research On Data Dissemination And Security Mechanism In Wireless Sensor Networks

Over-the-air data dissemination is an important means to manage and maintain wireless sensor networks(WSNs),and it is an indispensable service in a variety of application areas.This service can be implemented by using coding-based data dissemination protocols.The protocols must meet reliability,validity,scalability and security.The techniques of performance analysis and optimization are important means to guarantee the first three performances of the protocols.Furthermore,when wireless sensor networks are deployed in hostile environments,security is the first issue to be considered.In a WSN,sensor nodes are resource-constrained,wireless communication channel is open,and the deployment area of the network is special.Hence,we face new challenges in ensuring the security of the protocols.The techniques of performance analysis and optimization are the foundations for the research on the security of the protocols.Only by combining the two research fields can we ensure secure and reliable data dissemination in WSNs.This paper focuses on performance analysis methods,performance optimization techniques and security issues of coding-based data dissemination protocols.Our main contributions are summarized as below:1.A page dissemination time model was proposed to accurately forecast the time of page dissemination,which can be obtained by performing coding-based data dissemination protocol.The prediction results from existing analysis models have the larger deviation with real data from these protocols.In order to reduce the deviation,we presented a page dissemination time model,which was based on the formal descriptions of page encoding,page decoding and negotiation mechanism in great detail,to analyze qualitatively the time of page dissemination.Our model attached importance to the influence of encoding and decoding.The prediction results of our model fit well with the simulation results through the experimental analysis,and our model can describe accurately the time of page dissemination.Our model is suitable for the data dissemination protocols using random linear codes,and it is instructive for the design of the protocols,performance evaluations and performance optimizations.2.For obtaining the optimal page granularity,a Firefly Algorithm with feasibility-based rules was proposed to solve the minimization problem of the time of page dissemination,which has the features of strong nonlinearity and multi-constraints.Page granularity is the core parameter of coding-based data dissemination protocols,and its value will affect the performances of the protocols.Existing methods analyzed quantitatively page granularity based on the massive experiments,thus the time overheads and cost overheads of these methods are relatively high.To solve the problem,an objective function of minimizing the time of page dissemination,which has the features with strong nonlinearity and multi-constraints,was built based on our page dissemination time model.Then the process of optimizating the objective function was converted to the mobile process of firefly,according to the phototaxis of firefly.Furthermore,two important factors of brightness and attractiveness were analyzed,the location of the movement for the firefly was designed,and feasibility-based rules were established to select the better firefly.We searched the optimal parameter of page granularity through an iterative approach.Our algorithm can find effectively the global optimal solution,and it has the feature of fast convergence.By optimizing the key parameter of page granularity,it will in turn promote the performance analysis for the protocols.3.A low-overhead security enhancement scheme based on Hierarchical Hash Tree(HHT)was proposed to enhance the protection of DATA and COMMAND packets in coding-based data dissemination protocols,and it can resist pollution attacks.Pollution attack is the main threat to secure coding-based data dissemination protocol.Existing security schemes can resist pollution attacks,but the overheads of these schemes are too large.Excessive overhead will occupy seriously the limited resources of sensor node,thus this leads to the decline of the performances of these protocols.To solve the problem,Hierarchical Hash Tree was constructed according to the ideas of hierarchy and aggregation.Our scheme can resist pollution attack by combining HHT to one time digital signature scheme based on elliptic curve.The security of our scheme was proved formally.Furthermore,we analyzed mathematically the overheads of our scheme.By experimental analysis,our scheme can reduce markedly authentication overhead and communication overhead of code image.The reduction of that is more significant as the size of code image grows.4.A comprehensive evaluation algorithm,which was based on the mechanism of prosecution and trust,was proposed to identify malicious nodes having malicious behaviors in coding-based data dissemination,and it can resist internal attacks for WSNs.Malicious nodes are hidden inside a WSN and their malicious behaviors have more serious threats to secure data dissemination.Existing identification algorithms for malicious nodes were based on hierarchical wireless sensor networks,where the data flowed from sensor nodes to the base station.These algorithms are not applicable to data dissemination based on network coding.In the paper,we introduced the idea of prosecution and trust to build a weighted evaluation model,and we designed a comprehensive evaluation algorithm based on the model to identify malicious nodes in a WSN.The malicious behaviors were determined by detecting the abnormal conditions in the execution of coding-based data dissemination protocol.Then the suspected node was prosecuted,the influence of the prosecution for sensor nodes was quantified to trust weight,and malicious prosecutions were suppressed.In our algorithm,the trust weight of defendant node was compensated after malicious prosecution had been detected,and the restricted prosecution function was restored for defendant node.By experimental analysis,the proposed algorithm can accurately and quickly identify malicious nodes in the networks.Thus,we can build a secure and trusted network running environment to support over-the-air data dissemination.