基于信任评估的虚拟网故障诊断算法

网络虚拟化环境的新特点(网络拓扑动态变化、动态症状和故障关系、管理域独立和信息不准确性)对故障诊断提出了新的要求,提出一种改进的针对网络虚拟化环境下虚拟网和底层网故障诊断模型和诊断算法,通过虚拟网信任评估算法来提高故障诊断的准确率、降低误报率.仿真研究结果表明,在大规模和噪声大的虚拟网络环境中,提出的故障诊断算法取得了较好的诊断效果.

Virtual Network Fault Diagnosis Using Trust Evaluation

The flexible characteristics of virtual network bring to fault diagnosis new challenges, such as dynamic network topology, dynamic symptom-fault causality, Administrator domain independence and incomplete and inaccurate information. To address these challenges, symptom monitoring model, layer negotiation diagnosis model, fault diagnosis algorithm using fidelity evaluations and symptoms transfer module are presented. Trust evaluation module (including two-stage evaluation within virtual network and two-stage evaluations within substrate network) can improve diagnosis accuracy and reduce false positive. Simulation study shows that our algorithm can efficiently and accurately get a hypothesis set explaining symptoms observed, even when the observed symptoms are incomplete and inaccurate.