Virtualized network element fault analysis method and system based on multi-observation-dimension HMM

Abstract

The invention discloses a virtualized network element fault analysis method and system based on a multi-observation-dimension HMM, and relates to the technical field of information processing. According to the invention, the modeling is performed based on a multi-observation-dimension HMM model, the transition between fault states is considered when historical observation data is adopted to calculate the fault state probability, and multiple monitoring item observation data are integrated to calculate the joint probability, so the accuracy of an analysis result is further improved. According to the invention, automatic processing of the fault alarm is realized, the operation and maintenance cost of the clouded network platform is reduced, and the stability of the clouded network platform and the timeliness of fault response processing are improved.

Description

technical field [0001] The present invention relates to the technical field of information processing, in particular to a virtualized network element failure analysis method and system based on a multi-observation dimension HMM (Hidden Markov Model, Hidden Markov Model). Background technique [0002] see figure 1 As shown, Network Functions Virtualization (NFV, Network Functions Virtualization) provides a new way to design, deploy and manage network services. NFV integrates network functions such as Network Address Translation (NAT, Network Address Translation), firewall, intrusion detection, Functions such as domain name services and caching are separated from proprietary hardware and implemented in software. Each VNF (Virtualized Network Function, virtualized network function) includes multiple VNFCs (Virtualized Network Function Componet, virtualized network function components), and each VNFC is mapped to a VM (Virtual Machine, virtual machine). Because NFV requires a ...

AI Technical Summary

Problems solved by technology

However, the realization of virtualized network functions under MANO orchestration is composed of many microservices, and the location and analysis of faults are extremely complicated. It is by no means that simple rules manually configured can handle them correctly. Wrong processing rules may even introduce misoperations. risk, which will cause irreparable damage

For example, when a virtualized network element is heavily loaded or fails, the value of the monitoring item may be high. According to the manually defined processing rule, a pre-specified processing action is taken when the threshold exceeds the limit. The rule defines that when the threshold is exceeded, automatic Expansion or self-healing are obviously not suitable

Therefore, when artificially formulating automatic elasticity and automatic healing rules, the correctness and effectiveness of the rules cannot be guaranteed

Images