TCP RTT Estimation Filtering Delayed Acknowledgments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing TCP-based data networks face challenges in accurately estimating Round-Trip Time (RTT) due to issues like delayed acknowledgments, cumulative acknowledgments, and packet reordering, which can lead to inaccurate traffic management and performance metrics.
Innovation Solution
A system and method that employs a monitoring module to filter out the effects of delayed acknowledgments by comparing sequence numbers and acknowledgement numbers, using timers to associate timestamps, and applying filters to discard inflated RTT measurements, ensuring accurate RTT estimation by only considering sequentially related packets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If delayed-ack timers are used in TCP implementations, then network protocol compatibility is maintained, but RTT measurement precision deteriorates due to inflated measurements
Solution Approach 1:
The patent segments RTT measurements into two categories: filtered measurements (excluding delayed-ack influenced packets) and unfiltered measurements (including all packets). By segmenting the measurement population, the system maintains protocol compatibility with delayed-ack timers while obtaining a precise RTT estimate from the filtered subset that excludes inflated measurements.
Solution Approach 2:
The patent extracts and removes delayed-ack influenced packets from the RTT measurement process. By identifying packets affected by delayed-ack timers and excluding them from RTT calculations, the system extracts only the clean, uncorrupted measurements needed for accurate RTT estimation while leaving the delayed-ack mechanism intact for protocol compatibility.
2Device complexity
If simple RTT estimation methods are used, then device complexity is reduced, but measurement precision deteriorates due to packet loss and reordering effects
Solution Approach 1:
The patent implements feedback mechanisms where RTT measurements are continuously monitored and validated. Packets are tracked through sequence numbers and acknowledgment numbers, with feedback loops that identify and exclude measurements influenced by packet loss, reordering, or delayed-ack timers. This feedback-driven filtering maintains relatively simple device complexity while significantly improving measurement precision.
3Quantity of substance
If all received packets are used for RTT estimation, then measurement data quantity increases, but measurement precision deteriorates due to inclusion of inflated RTT values
Solution Approach 1:
The patent applies local quality filtering where different quality standards are applied to different packets based on their characteristics. Packets are evaluated individually for signs of delayed-ack influence, packet loss, or reordering, and only high-quality measurements meeting specific criteria are included in RTT estimation. This selective approach maintains sufficient measurement data quantity while ensuring high precision by excluding degraded measurements.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system and method for estimation of round trip times (RTTs) within a TCP based data network is described. The system provides for a passive monitoring of the network round-trip latency by simply observing existing TCP connections in operation. In this way, it is possible to measure RTT out to a remote site without having to send any traffic to that site with the result that there is no impact on the network, and it is possible to estimate what production traffic is experiencing.