Network Performance Estimation via Reference Packets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining network performance across multiple wide area network (WAN) interfaces are time-consuming and complex, especially when dealing with aggregated end-to-end connections, and can be impacted by traffic passing through independently connected connections.
Innovation Solution
The method involves selecting network interfaces at both nodes, transmitting reference packets, receiving arrival information, and estimating network performance based on this information, with the option to establish an aggregated end-to-end connection and use policies to optimize interface selection and performance measurement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If network performance testing is performed across multiple WAN interfaces and end-to-end connections, then comprehensive network performance information is obtained, but the testing becomes time-consuming and complex
Solution Approach 1:
The patent segments the network performance testing by establishing multiple independent end-to-end connections between different WAN interfaces of the first node and second node. Each connection is tested separately using reference packets, allowing parallel testing without interference. This segmentation enables comprehensive coverage of all M×N interface combinations while maintaining manageable testing complexity and reduced overall testing time.
2Measurement precision
If traffic is transmitted through end-to-end connections during performance testing, then network performance data is collected, but the traffic impacts network performance of other end-to-end connections
Solution Approach 1:
The patent extracts the performance measurement function from regular data traffic by using dedicated reference packets for testing. These reference packets are specifically designed for performance measurement purposes and are transmitted through established end-to-end connections. This separation allows accurate performance collection without allowing test traffic to interfere with or degrade the network performance of other active connections.
3Loss of information
If all network performance information is displayed simultaneously, then complete information is provided, but it becomes difficult to display and interpret
Solution Approach 1:
The patent merges multiple end-to-end connections into aggregated end-to-end connections, where performance information from individual connections is combined and displayed collectively. This aggregation reduces the complexity of information display while maintaining completeness, as users can view overall network performance across all WAN interface combinations without being overwhelmed by individual connection details. The system preserves the ability to access individual connection data when needed.
Data Source
AI summary
A method carried out by a first node for estimating network performance by selecting at least one network interface of the first node and at least one network interface of a second node and then transmitting reference packets to the second node. The reference packets are transmitted through the selected network interfaces of the first node to the selected network interfaces of the second node. After receiving the reference packets, the second node generates arrival information and sends the arrival information to the first node. The first utilizes the arrival information for estimating the network performance. The network interfaces are selected according to a policy and end-to-end connection(s) may be established between selected network interfaces of the first node and selected network interfaces of the second node. The end-to-end connection(s) may be aggregated to form an aggregated end-to-end connection. Alternatively, the network interfaces are selected according to an end-to-end connection profile or an aggregated end-to-end connection profile.