Packet Sampling for Granular Flow Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for performance measurement in packet-switched communication networks face scalability issues when measuring a large number of packet flows, requiring complex and costly filters at network nodes, and are unable to provide granular measurements for specific portions of multipoint packet flows.
Innovation Solution
A method that uses a hash function to select sample packets from multipoint packet flows at measurement points, with a management server identifying clusters and performing performance measurements on packet sub-flows based on packet content, allowing for off-line filtering and granular measurements without complex filters at network nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex filters are implemented at network nodes to identify and measure each packet flow, then measurement precision is improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent extracts the complex filtering and measurement logic from network nodes and relocates it to a centralized analysis server. Measurement points at network nodes only perform simple packet sampling and forwarding, while the analysis server handles the complex flow identification and performance measurement, thereby reducing device complexity at nodes while maintaining measurement precision.
Solution Approach 2:
The patent introduces measurement points as intermediary components between network nodes and the analysis server. These measurement points capture packet samples and forward them to the analysis server, which then performs the complex measurement operations. This intermediary layer allows simple node devices to achieve precise measurements through the centralized analysis server.
2Quantity of substance
If multiple complex filters are deployed at network nodes to measure numerous packet flows, then measurement coverage is improved, but scalability deteriorates due to high implementation cost
Solution Approach 1:
The analysis server is designed as a universal platform that can measure multiple packet flows simultaneously using a single centralized system. Instead of deploying separate filter complexes at each node for each flow, the analysis server handles identification and measurement of numerous flows through software-based filtering, significantly improving scalability and reducing implementation costs.
Solution Approach 2:
The patent uses packet sampling where measurement points create copies of packets for analysis while the original packets continue through the network. This copying approach allows the analysis server to examine packet content and perform measurements without affecting network performance, enabling comprehensive measurement of multiple flows with minimal impact on productivity.
3Measurement precision
If complex filters are implemented at measurement points to identify specific packet flows, then measurement precision is improved, but ease of operation deteriorates due to configuration complexity
Solution Approach 1:
The patent extracts the complex flow identification logic from measurement point configurations and relocates it to the analysis server. Measurement points only need simple configuration to capture packets and forward them to the analysis server, which then performs sophisticated flow identification based on packet content, significantly simplifying the configuration process while maintaining high identification accuracy.
Solution Approach 2:
The analysis server automatically performs flow identification and measurement without requiring complex manual configuration at measurement points. The system self-configures by analyzing packet content and automatically identifying flow patterns, reducing the operational burden on network operators while maintaining precise measurement capabilities.
Data Source
AI summary
A method for performing a performance measurement in a communication network. Each measurement point in the network identifies packets of a multipoint packet flow and selects therefrom a number of samples, based on the value of a sampling signature calculated by applying a hash function to a bit mask in each identified packet. For each sample, a performance parameter and the packet's content are provided to a management server. The management server identifies a cluster of measurement points such that each identified packet of the multipoint packet flow received by a cluster's input measurement point is also received at a cluster's output measurement point. Amongst the performance parameters provided by the cluster's measurement points, the performance parameters relating to samples belonging to a certain packet sub-flow are identified, based on the packet's content. Then, a performance measurement is performed on the packet sub-flow.