Network Model Aware Diagnosis for Root Cause Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current network management systems struggle with efficiently performing root cause analysis in complex networks with frequent changes and multiple concurrent faults, as existing techniques either fail to handle frequent updates or are limited to single-issue scenarios.

Innovation Solution

The implementation of a Network Model Aware Diagnosis technique that uses a combination of element models, service models, and multi-layer models to capture dependencies and cause-and-effect relationships across resources, services, and service events, allowing for node-level, service-level, and multi-layer root cause analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional network management systems are used to perform root cause analysis, then the system can identify network issues, but the analysis becomes inefficient and inaccurate in complex networks with frequent changes and multiple concurrent faults

Engineering Contradiction:
Improveroot cause analysis accuracyVSAvoidanalysis time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system pre-generates multiple hypothetical root cause scenarios and their corresponding expected network states before actual failures occur. When a failure is detected, the system compares the actual network state against these pre-computed scenarios to rapidly identify the root cause, eliminating the need for time-consuming post-failure analysis

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediary layer that maintains a unified graph model representing the entire network topology and state. This intermediary model acts as a mediator between raw network data and root cause analysis, enabling efficient querying and comparison to identify failures across complex multi-layer networks

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If the network model includes detailed dependencies and cause-and-effect relationships across multiple layers, then the root cause analysis becomes more accurate, but the model complexity and computational overhead increase

Engineering Contradiction:
Improvefailure identification accuracyVSAvoidmodel complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system segments the complex network model into hierarchical layers (physical layer, logical layer, service layer) while maintaining a unified graph structure. Each layer can be independently modeled and analyzed, reducing the complexity of managing full-network dependencies while preserving accuracy through layered composition

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The unified graph model serves multiple functions simultaneously: it represents network topology, captures dependency relationships, stores state information, and enables various types of analysis (failure detection, root cause identification, impact assessment). This multi-functionality reduces the need for separate complex models for each purpose

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3823215B1Network model aware diagnosis of a network
Publication Date: 2024.11.13 JUNIPER NETWORKS INC
  • EP3823215B1 patent drawingFigure 1
  • EP3823215B1 patent drawingFigure 2
  • EP3823215B1 patent drawingFigure 3~4

AI summary

Management of network devices using network aware diagnosis techniques. An example method to monitor a network includes constructing, by a controller device that manages a plurality of network devices, a deployed graph model comprising a plurality of models representative of services operating on the network devices within the network. The method also includes receiving an alert, by the controller device, that a service operating on one of the network devices is experiencing a failure. Additionally, the method includes traversing, by the controller device, modeled dependencies within the deployed graph model to perform root cause analysis of the service that is experiencing the failure to determine a resource-level event causing the failure. The method includes. based on the traversing, providing, by the controller device, a report indicating the root cause of the failure.