选择性转发攻击下的城市小区域车联网级联失效机理及防御策略

The internet of vehicle in small regions of the city has the features of scale-free network, so the intentional attack is the gravest threat to the security of the network. If node failures come from intentional attack based on the feature of degree distribution inhomogeneity, the failure is a kind of dynamic effect of node failure, will instantly spread all over the network, which causing global network paralysis, becoming the application bottleneck in the network. Now, the difficult of selective forward attack detecting and the complexity of cascading failure analyzing, pose a great challenge for internet of vehicle network. Therefore, this project studies the malicious nodes which selectively drop legal data packets from time to time. Using the Gauss distribution and the Bayesian decision to analyze the behavior of packet loss changes, the measurement criteria for the selective forwarding attack node is established. Then the cascading failure effect on the network traffic variation based on generating function method is analyzed, the trigger critical traffic distribution of network large-scale cascading failure under the selective forwarding attack is obtained. Moreover, by introducing the chance of entropy into the construct process of the path traffic allocation model, and using multi-paths that work together to forwarding data packets, the cascading failure defense strategy is further proposed. It reduces the malicious nodes involved in the possibility of forwarding data packets, in the meantime, it ensures the reasonable traffic distribution for each path in the network. Finally, the completion of the project will reveal the cascading failure mechanism under selective forwarding attack for the internet of vehicle in small regions of the city, enrich the cascading failure defense approach, and has an important scientific and practical value in improving the security of the internet of vehicle and implementing the strategy of the “the architecture of Internet of things in the space, the air and the land”.

城市小区域的车联网呈现出无标度网络特征，使得网络安全面临严重威胁，若攻击者结合网络度分布不均匀特征发起攻击，更会造成节点失效的动态效应，导致全局网络瘫痪。目前，选择性转发攻击的隐蔽性和级联失效的复杂性，使得车联网安全研究充满挑战。为此，本项目面向恶意节点丢弃合法数据包的选择性转发攻击，利用高斯分布和贝叶斯判决分析节点丢包行为，建立选择性转发攻击节点的检测标准；基于概率母函数法研究攻击下网络流量变化对级联失效规模影响规律，获取攻击下触发网络大规模级联失效的临界流量分布；引入机会熵构建网络路径流量分配模型，通过多条数据包传送路径的协同工作设计级联失效防御策略，减少恶意节点参与数据包转发的可能性同时保证网络各条路径流量分配的合理性。项目的完成将揭示攻击下城市小区域车联网级联失效机理，丰富级联失效防御方法，对提升车联网安全和贯彻“空天地智慧物联网架构”战略思想具有科学意义和实际价值。