Network Throughput Measurement via Packet Activity Symbols
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network operators face challenges in accurately measuring network throughput due to encrypted and obfuscated data flows, which hinder the evaluation of per-object download times and overall user experience, especially with the increasing use of encrypted protocols like QUIC that combine multiple flows into a single connection.
Innovation Solution
The solution involves periodically sampling network packet flows, generating activity records with symbols indicating active or idle periods, and inferring throughput suitability without interpreting payload contents, allowing network operators to distinguish suitable data flows for performance evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If encrypted protocols like QUIC are used to combine multiple flows into a single connection, then network security and protocol efficiency are improved, but the ability to measure per-object download times and evaluate user experience deteriorates
Solution Approach 1:
The patent segments the encrypted data stream by identifying and separating individual object downloads within the combined QUIC flow. By analyzing packet patterns, timing intervals, and flow characteristics, the system divides the encrypted stream into distinguishable object transfer segments, enabling per-object performance measurement while maintaining encryption integrity.
Solution Approach 2:
The patent introduces an intermediary measurement system that operates at the network layer rather than requiring payload inspection. This intermediary analyzes metadata, packet timing, and flow characteristics to infer object download performance without decrypting the encrypted stream, thus preserving security while enabling measurement.
2Measurement precision
If packet payload contents are inspected to identify suitable data flows for throughput measurement, then measurement accuracy is improved, but processing complexity and computational overhead increase
Solution Approach 1:
The patent extracts only the necessary metadata and flow characteristics from packets required for throughput measurement, such as packet timing, size, and flow identification fields. By taking out only these essential elements rather than fully inspecting or decrypting payload contents, the system achieves accurate measurement with reduced processing complexity.
Solution Approach 2:
The patent applies partial inspection by examining only sufficient portions of packet data needed to identify suitable flows for measurement. Rather than performing complete payload analysis, the system uses targeted inspection of flow characteristics and packet patterns to determine measurement suitability, reducing computational overhead while maintaining accuracy.
Data Source
AI summary
Aspects of the subject disclosure may include, for example, identifying a flow of data packets between first and second network addresses of a network, with each packet including respective header and payload portions. The identified flow of data packets is monitored over a number of sample periods to obtain a number of monitored results. A data-flow activity record is generated, having a number of symbols corresponding to the number of monitored results, the symbols including an active symbol value indicative of a presence of an exchange of data and an idle symbol value indicative of an absence of an exchange of data. A suitability of the identified data flow is inferred for estimating a throughput of the network according to the data-flow activity record without interpreting contents of each respective packet payload portion. Other embodiments are disclosed.