Network Congestion Management via Forward Path Latency Variation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current congestion management methods in data communication networks rely on feedback from network devices, which can be cumbersome and fail to accurately inform end-to-end latency, particularly distinguishing between forward and reverse path congestion.
Innovation Solution
A method and system that determine congestion state based on the variation of forward path latency by embedding timestamps in packets, allowing for congestion management without requiring synchronized clocks or intermediate network hardware, and enabling congestion notification independent of communication protocols.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If congestion management relies on feedback from network devices (switches/routers), then congestion detection capability is improved, but device complexity and configuration requirements increase
Solution Approach 1:
The patent extracts congestion detection functionality from intermediate network devices and relocates it to end nodes. By having end nodes measure forward path latency themselves using timestamps, the system eliminates the need for ECN marks, QCN feedback, or other intermediate device-based detection mechanisms, thereby reducing device complexity while maintaining detection accuracy
Solution Approach 2:
End nodes perform self-service congestion detection by measuring their own forward path latency using embedded timestamps. This eliminates reliance on intermediate network devices to provide congestion feedback, reducing both device complexity and configuration requirements while enabling accurate congestion state determination
2Measurement precision
If RTT-based congestion detection is used, then congestion measurement capability is improved, but measurement precision deteriorates due to lumping queueing in both directions
Solution Approach 1:
The patent segments the round-trip measurement into separate forward path and reverse path components. By measuring only the forward path latency using timestamps embedded in data packets traveling from sender to receiver, the system isolates forward path congestion information from reverse path ACK queueing, thereby preventing information loss and improving measurement precision
Solution Approach 2:
The patent introduces timestamps as an intermediary mechanism to enable precise forward path latency measurement. These timestamps are embedded in data packets and processed by the receiver to calculate one-way delay, serving as a mediator that separates forward path measurement from the confounding reverse path effects inherent in RTT-based methods
3Loss of information
If ECN marks are used for congestion notification, then congestion signal transmission is improved, but latency information accuracy deteriorates as queueing delay is not directly reflected
Solution Approach 1:
The patent replaces the mechanical ECN marking system with a direct measurement approach. Instead of relying on queue occupancy thresholds that indirectly indicate congestion, the system directly measures forward path latency using timestamps, providing precise end-to-end latency information that accurately reflects actual queueing delay without the information loss inherent in ECN marking
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for managing network congestion is provided. The method comprises: receiving, at a receiver, a packet comprising a timestamp provided by a first clock of a sender; deriving, by the receiver, a latency value based at least in part on the timestamp provided by the first clock and a receipt time provided by a second clock of the receiver; determining a latency change by comparing the latency value with a previous latency value; and determining a state of network congestion based at least in part on the latency change.