Router Packet Loss Detection Using CoS-Based Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing packet loss detection methods in virtual private networks, such as VPLS or VLL, using pseudo-wires, face inaccuracies due to packet disorder caused by different Class of Service (CoS) values and the inefficiency of in-line real-time flow monitoring, which results in inaccurate measurement results and excessive bandwidth usage.
Innovation Solution
A packet loss detecting method and apparatus that count packets according to their CoS values, generate and send packet loss detection packets with corresponding CoS values, and measure packet loss based on locally recorded and received count values, ensuring accurate detection without packet disorder and minimizing bandwidth usage by deploying measurement flows corresponding to the number of CoS values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If packet loss detection packets are sent with highest CoS value to ensure priority transmission, then transmission speed is improved, but packet disorder occurs and measurement accuracy deteriorates
Solution Approach 1:
The patent segments packet loss detection by creating separate detection flows for different CoS values. Instead of using a single high-priority detection packet that causes disorder, the system divides detection into multiple independent flows, each matching a specific CoS level, thereby preventing packet disorder while maintaining accurate measurement for each service class.
Solution Approach 2:
The patent changes the CoS parameter of detection packets to match the CoS value of the service flow being measured. Rather than always using the highest CoS value for detection packets, the system dynamically adjusts the detection packet's CoS parameter to correspond with the target service flow, ensuring accurate measurement without causing packet disorder.
2Measurement precision
If in-line real-time flow monitoring is deployed for each service flow to achieve accurate packet loss detection, then measurement precision is improved, but bandwidth consumption and device complexity increase excessively
Solution Approach 1:
The patent makes detection packets multi-functional by enabling them to carry both detection functionality and service flow identification (through matching CoS values). A single detection packet structure serves multiple purposes: it detects packet loss and simultaneously identifies which service flow is being measured, eliminating the need for separate monitoring infrastructure for each service flow.
Solution Approach 2:
The patent uses detection packets that copy the essential characteristics (CoS value) of the service flows being measured. Instead of deploying complex in-line monitoring equipment for each flow, the system creates simplified detection packet copies that traverse the network and provide measurement data, significantly reducing device complexity while maintaining measurement precision.
Data Source
AI summary
A packet loss detecting method includes: a first router counts received packets and sent packets respectively according to a class of service value of a packet in each tunnel; the first router generates a first packet loss detection packet corresponding to the class of service value, sends the first packet loss detection packet to a second router at a peer end, and receives a first packet loss response packet returned by the second router; and the first router corresponds a class of service value, and measures a packet loss value according to a locally recorded count value and a count value carried in the first packet loss response packet. With the technical solution, packet disorder does not occur when packet loss measurement is performed for a pseudo-wire with multiple classes of service. In addition, traffic detection occupies a smaller bandwidth.