Packet Sampling Signature for Network Performance Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing packet-switched communication networks face challenges in efficiently measuring performance metrics like packet loss, one-way delay, and jitter due to high data generation and processing requirements, and are prone to reception sequence errors and unpredictable sampling rates.
Innovation Solution
Implementing a method where each measurement point calculates a sampling signature for received packets and dynamically adjusts the signature length based on the number of selected samples to control the sampling rate, ensuring the number of measurement samples does not exceed storage and computational capabilities, thereby reducing data processing and improving robustness against errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If each single packet is identified and analytics data is generated for each packet, then measurement precision is improved, but computation effort and bandwidth consumption increase remarkably
Solution Approach 1:
The patent segments the packet identification process by using only specific bits from the packet header (source IP, destination IP, source port, destination port) to create a sampling signature, rather than processing the entire packet. This segmentation reduces the computation effort while maintaining the ability to identify and measure packet loss accurately.
Solution Approach 2:
The patent extracts only the necessary identifying bits from each packet to create a compact sampling signature. By taking out only the essential identification elements (40 bits from header fields) and discarding the rest, the system achieves accurate packet loss measurement without the remarkable computation effort required for full packet processing.
2Adaptability or versatility
If a hash function is used to identify each packet, then packet identification capability is improved, but reception sequence errors and collisions occur reducing reliability
Solution Approach 1:
The patent changes the parameters of identification by using a fixed-length bit mask (40 bits) selected from specific packet header fields, rather than using a full hash function. This parameter change reduces collisions and reception sequence errors while maintaining packet identification capability, thereby improving measurement reliability.
3Measurement precision
If the sampling rate is increased to capture more packets, then measurement precision is improved, but data processing burden increases
Solution Approach 1:
The patent applies local quality by creating a compact sampling signature using only specific bits from the packet header rather than processing the entire packet. This localized approach allows for higher sampling rates with reduced processing burden, as each signature requires minimal computation and storage resources.
4Loss of information
If all analytics data from all observation points is gathered and processed, then comprehensive network analysis is improved, but bandwidth consumption and processing time increase
Solution Approach 1:
The patent creates a simplified copy of packet identification information in the form of a sampling signature. Instead of gathering and processing complete analytics data from all observation points, the system uses these compact signatures to identify and correlate packets across the network, significantly reducing bandwidth consumption and processing time while maintaining analytical completeness.
Data Source
AI summary
It is disclosed a method for performing a performance measurement on a packet flow transmitted along a path through a packet switched communication network. Two or more measurement points are implemented on the path. Each measurement point calculates a sampling signature for each received packet by applying a hash function to a mask of bits of the packet. Then, it selects a number of measurement samples amongst the received packets, the measurement samples being selected as those packets whose sampling signatures comprise a portion of length S equal to a predefined sampling value. While performing the selection, the measurement point counts the number of selected measurement samples and retroactively adjusts the length based on this number. Then, the measurement point provides measurement parameters relating to the selected measurement samples.