Application-Specific High Frequency Passive Probing for SD-WAN
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Quantifying application experience in software-defined wide area networks (SD-WANs) has become increasingly difficult due to the shift from Transmission Control Protocol (TCP) to User Datagram Protocol (UDP)-based traffic, as existing methods rely on control signaling which is less effective for UDP-based protocols like HTTP/3 and QUIC, which obscure retransmits and congestion metrics.
Innovation Solution
The implementation of application-specific high frequency passive probing, where packet maps are generated as multi-dimensional histograms indexed by packet properties and time, allowing for the passive inference of network characteristics and application experience, even for fully encrypted traffic, by comparing packet maps between networking devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If control signaling with TCP traffic is used to capture metrics, then metrics regarding retransmissions and response times can be captured, but the method becomes ineffective for UDP-based protocols like HTTP/3 and QUIC which obscure retransmits and congestion metrics
Solution Approach 1:
The patent creates packet maps that serve as a universal measurement mechanism applicable to both TCP and UDP protocols. Instead of relying on protocol-specific control signaling, the system uses a unified packet inspection approach that generates multi-dimensional histograms of packet properties, making it versatile across different transport protocols while maintaining measurement effectiveness
Solution Approach 2:
The patent introduces packet maps as an intermediary representation layer between the network traffic and the measurement process. These packet maps serve as a mediator that translates raw packet data into meaningful metrics without requiring protocol-specific control signaling, enabling effective measurement for both TCP and UDP traffic through a common interface
2Productivity
If passive probing is used to infer network characteristics, then network metrics can be measured without active traffic generation, but the precision and accuracy of inferred metrics may be compromised
Solution Approach 1:
The patent transforms packet data into multi-dimensional histograms (packet maps) with dimensions including packet size, inter-arrival time, and other packet properties. This dimensional transformation enriches the passive probing data structure, allowing more precise inference of network metrics from passive observations by analyzing patterns across multiple dimensions simultaneously
Solution Approach 2:
The patent performs preliminary processing of packet data by generating packet maps that organize and pre-process traffic characteristics before metric inference. This preliminary structuring of data into multi-dimensional histograms prepares the information in advance, enabling more accurate passive measurement without requiring active probing
Data Source
AI summary
In one embodiment, a first networking device in a network coordinates, with a second networking device in the network, capture of packet maps for a traffic flow in the network associated with a particular application. The packet maps comprise multi-dimensional histograms indexed by identified properties of packets of the traffic flow and time. The first networking device inspects packets of the traffic flow, to identify properties of packets of the traffic flow. The first networking device generates a first packet map for the traffic flow based on the properties of the packets of the traffic flow identified by the first networking device. The first networking device causes a comparison between the first packet map and a second packet map generated by the second networking device to be used as a measure of application experience for the particular application.