基于资源特征的虚拟网服务故障诊断算法

VIRTUAL NETWORK SERVICE FAULT DIAGNOSIS ALGORITHM BASED ON RESOURCE CHARACTERISTICS

  • 摘要: 虚拟化环境下,网络的多层复杂性和信息不准确性导致虚拟网服务故障诊断算法的准确率较低。为解决此问题,基于服务、虚拟网资源、底层网络资源的特征和关联关系,提出一种基于资源特征的虚拟网服务故障诊断算法。根据服务承载和资源映射关系,构建疑似虚拟故障集合。根据症状数量、虚拟节点数量、负症状与底层节点的概率模型,分析故障节点的可信度并动态构建故障传播模型。通过分析疑似故障集合对负症状的解释能力,对虚拟网服务故障进行定位。从简单网络环境和复杂网络环境两个方面,验证该算法较好地提升故障诊断的准确率,降低故障诊断的误报率和诊断时长。

     

    Abstract: In the virtualized environment, the multi-layer complexity of the network and the inaccuracy of information lead to the low accuracy of the virtual network service fault diagnosis algorithm. In order to solve this problem, based on the characteristics and associations of services, virtual network resources and substrate network resources, a fault diagnosis algorithm for virtual network services based on resource characteristics is proposed. According to the service bearer and resource mapping relationship, a set of suspected virtual faults was constructed. According to the number of symptoms, the number of virtual nodes, the probability model of negative symptoms and substrate nodes, the reliability of faulty nodes was analyzed and a fault propagation model was dynamically constructed. The virtual network service fault was located by analyzing the ability of the suspected fault set to explain the negative symptoms. From the two aspects of simple network environment and complex network environment, it is verified that this algorithm can improve the accuracy of fault diagnosis, and reduce the false alarm rate and diagnosis time of fault diagnosis.

     

/

返回文章
返回