网络虚拟化环境下,底层网络的透明性造成虚拟网提供商不能诊断所有的虚拟网服务故障.为解决此问题,提出了基于映射关系的虚拟网服务故障传播模型.针对故障传播模型中故障集与症状集较大、网络环境动态和噪声大而导致的已有诊断算法误报率高、时间复杂度高的问题,基于网络虚拟化环境下症状内在相关性特点,提出了一种新的基于症状内在相关性的虚拟网服务故障诊断算法SFDoIC(service fault diagnosis algorithm based on inherent correlation among symptoms).仿真实验结果表明,SFDoIC算法能够很好地解决底层网络透明性造成的虚拟网服务故障难以定位的问题.SFDoIC算法可以有效地降低诊断算法的误报率,显著缩短诊断算法的运行时间.
The virtual network provider(VNP) cannot diagnose all service faults of virtual networks,because the substrate network is transparent for VNP within the network virtualization environment.To solve this problem,the paper presents a service fault propagation model based on mapping relationships.In terms of the large fault set,the large symptom set,and noisy and dynamic environments,which result in the higher false positive rate and the longer running time of existing fault diagnosis algorithms,a service fault diagnosis algorithm based on inherent correlation among symptoms(SFDoIC),is proposed.Simulation results show that algorithm SFDoIC can solve the difficult problems in fault diagnosis that are caused by the transparency of the substrate network for VNP,effectively reducing the false positive rate and decreasing running time.