Network Fault Locating via Flooded KPI Packets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network fault locating methods are inefficient due to difficulties in collecting fault information from all network devices, especially when some devices do not support network management, leading to slow fault analysis and potential network flapping and service interruptions.
Innovation Solution
A fault locating method that involves flooding fault information packets with key performance indicators (KPIs) across network devices, allowing for quick identification of faulty devices based on KPI change rates and centrality analysis using an adjacency matrix, thereby reducing fault locating time and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Telnet is used to log in to different network devices to collect fault information, then fault information can be collected from devices, but the fault information collection efficiency is low and time-consuming
Solution Approach 1:
The patent uses a flooding mechanism where fault information packets are copied and transmitted to all network devices simultaneously. Instead of manually logging into each device, the system creates copies of the fault inquiry packet and distributes them across the network, allowing parallel collection of fault information from all devices at once, thus dramatically reducing fault locating time while maintaining complete information collection
Solution Approach 2:
The patent implements preliminary action by pre-configuring fault information collection mechanisms and KPI monitoring on all network devices before faults occur. The devices are prepared to immediately respond to flooding packets with relevant fault data, eliminating the need for manual login and real-time information gathering during actual fault events
2Measurement precision
If manual troubleshooting is performed on massive information from many network devices, then fault analysis can be conducted, but the analysis efficiency is low and requires rich device operation experience
Solution Approach 1:
The patent extracts only the most relevant fault information by implementing KPI (Key Performance Indicator) filtering mechanisms. Instead of analyzing all massive data from network devices, the system identifies and extracts critical KPIs that are most indicative of faults, such as packet loss rates, latency metrics, and error counts. This extraction process automates the analysis by focusing computational resources on the most diagnostic information, improving both efficiency and accuracy while reducing the need for manual expert analysis
Solution Approach 2:
The patent implements feedback mechanisms where collected fault information and KPIs are continuously analyzed and fed back to refine the fault locating process. The system uses the extracted KPI data to automatically update fault probability assessments and adjust subsequent information collection strategies, creating a closed-loop system that improves analysis efficiency through automated iterative refinement rather than manual troubleshooting
3Measurement precision
If fault information is collected from all network devices in the entire network, then comprehensive fault analysis is possible, but some devices do not support network management and cannot be accessed
Solution Approach 1:
The patent implements universality by designing a flooding-based fault information collection mechanism that works across diverse network devices regardless of their management capabilities. The system uses standardized protocol packets that can be processed by any network device, whether managed or unmanaged. This universal approach allows the system to collect fault information from all devices in the network using a single mechanism, achieving complete information coverage without requiring device-specific access methods
Data Source
AI summary
A fault locating method includes: receiving, by a first network device, a fault information packet flooded by another network device in a network, where each fault information packet includes statistical information about an interior gateway protocol packet of the network device sending the fault information packet, and the statistical information of each network device includes a statistical result on one or more key performance indicators KPIs of the network device; and determining, based on statistical information of the first network device and the statistical information of the another network device, a network device on which a fault occurs in the network. According to the method, the fault information packet flooded by the network device is received, so that information required for fault locating, that is, the KPI of the network device, can be quickly collected. This facilitates a fault locating process and reduces a fault locating time.