Ethernet Link Traffic Analysis via Packet Loss Rate
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Ethernet traffic statistics and analysis methods cannot accurately reflect the current state of a link or quickly locate fault positions, as they only analyze port-level data and fail to account for packet loss on links, leading to inaccurate network traffic distribution and fault evaluation.
Innovation Solution
An Ethernet traffic statistics and analysis method that involves two traffic analyzed nodes connected to a link, where a traffic analyzing node counts and constructs an analytic diagram based on the rate of lost packets passing through the link, using port traffic information reported at regular intervals to accurately reflect link traffic and identify faults.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If port-level traffic information is collected and analyzed, then network traffic monitoring is implemented, but link traffic state and fault location cannot be accurately reflected
Solution Approach 1:
The patent segments the network monitoring function into two parts: port-level monitoring (existing) and link-level monitoring (new). By dividing the monitoring scope from just ports to include the links connecting them, the system can now separately analyze packet loss information on each link using the formula: link packet loss = sum of sent packets on both ports - sum of received packets on both ports. This segmentation resolves the contradiction by enabling precise link-level measurement while preserving port-level monitoring capabilities.
Solution Approach 2:
The patent adds a new dimension to traffic monitoring by introducing link-level analysis alongside existing port-level monitoring. Instead of only monitoring traffic entering/leaving individual ports, the system now monitors the traffic flow across the link connecting two ports. This dimensional expansion allows simultaneous observation of both port traffic statistics and link packet loss, resolving the information loss problem while maintaining measurement precision.
2Productivity
If only node port traffic information is counted, then statistical analysis is simplified, but fault location and network abnormality detection are delayed
Solution Approach 1:
The patent makes the traffic analyzing node multi-functional by enabling it to perform both port-level traffic analysis (existing function) and link-level packet loss analysis (new function). The analyzing node collects port traffic information from multiple nodes and simultaneously calculates link packet loss by comparing sent and received packet counts. This universal approach allows the same node to detect both normal traffic variations and abnormal packet loss conditions, improving fault detection speed without requiring separate dedicated devices.
Solution Approach 2:
The patent introduces link packet loss rate as an intermediary metric that mediates between port-level traffic data and fault detection. Instead of directly analyzing complex port traffic patterns to detect faults, the system uses the intermediary calculation of link packet loss (sent packets - received packets) to identify abnormalities. This intermediary approach simplifies the analysis process while improving fault detection speed, as packet loss on a link is a clear indicator of link faults rather than port configuration issues.
3Reliability
If traffic rate decrease is used as fault indicator, then fault detection is possible, but false positives occur due to normal traffic flow variations
Solution Approach 1:
The patent implements feedback by continuously monitoring link packet loss rates and comparing them against threshold values. The system collects port traffic information at regular intervals, calculates link packet loss using the feedback formula (sent - received), and triggers fault alerts only when packet loss exceeds the threshold. This feedback mechanism distinguishes between normal traffic fluctuations (where packet loss remains low) and actual link faults (where packet loss increases significantly), improving detection reliability while maintaining precise link state measurement.
Data Source
Figure 1~3
Figure 4~5
Figure 6~7
AI summary
An Ethernet traffic statistics and analysis method comprises: according to two pieces of port traffic information reported continuously by each of two traffic analyzed nodes which are connected to any one link and a reporting time interval, a traffic analyzing node counting traffic of packets passing through the link, traffic of lost packets and corresponding rates, and constructing an analytic diagram according to the rate of the traffic of the lost packets counted every time. An Ethernet traffic statistics and analysis system is also provided. The solution performs statistics and analysis on the traffic on the link in the Ethernet, and reflects the current state of the link accurately and visually. When a fault occurs or a packet is lost on the link, the fault position can be quickly located, and the efficiency of management and maintenance can be increased.