Dynamic Congestion Control via Adaptive Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current window-based congestion control schemes in communication networks suffer from excessive throughput oscillations and fixed bandwidth sacrifices, leading to inefficiencies in packet forwarding due to stochastic active queue management (AQM) and predefined capacity thresholds.
Innovation Solution
A networking endpoint with a source congestion control module that adapts packet transmission based on congestion signals, using multiple probing schemes (aggressive, slower, and back-off) and load measurement with dynamic thresholds to avoid oscillations and optimize bandwidth usage, incorporating an exponential weighted moving average for load estimation and explicit congestion notification (ECN) for proactive flow control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If a fixed portion of link capacity is sacrificed to absorb oscillations (HULL architecture), then queue oscillations are reduced, but throughput efficiency decreases due to fixed bandwidth loss
Solution Approach 1:
The patent implements dynamic threshold adjustment where the congestion threshold is no longer fixed but adapts based on current network conditions and traffic patterns. The forwarding device dynamically modifies the threshold value to match actual oscillation characteristics, allowing the system to maintain stability only when necessary while maximizing throughput during normal operation. This resolves the contradiction by making the bandwidth sacrifice conditional rather than permanent.
Solution Approach 2:
The system changes the threshold parameter dynamically based on measured queue behavior and oscillation patterns. When oscillations are detected, the threshold is adjusted to absorb them; when stable, the threshold is raised to maximize throughput. This parameter adaptation allows the system to transition between stability-focused and efficiency-focused modes, resolving the fixed trade-off.
2Stability of the object's composition
If stochastic AQM schemes are used to reduce queue oscillations, then some stability is achieved, but unpredictable delays occur due to stochastic packet dropping
Solution Approach 1:
The patent implements feedback mechanisms where the forwarding device monitors queue behavior, detects oscillations, and adjusts thresholds based on this feedback. The system continuously adapts to actual network conditions rather than using predetermined stochastic dropping patterns. This feedback-driven approach maintains stability while preserving predictability because adjustments are made responsively rather than randomly.
Solution Approach 2:
The network system performs self-diagnosis and self-adjustment by monitoring its own queue behavior and automatically modifying thresholds to correct oscillations. This eliminates the need for external control or predetermined stochastic patterns, allowing the system to maintain both stability and predictability through autonomous adaptation to actual conditions.
3Loss of time
If zero-queuing delay is pursued for latency-sensitive applications, then latency is minimized, but throughput decreases due to lack of queuing buffer
Solution Approach 1:
The patent implements dynamic threshold adjustment that adapts to traffic patterns and oscillation characteristics. During periods of stable traffic, the threshold is set high to maximize throughput with minimal queuing. When oscillations are detected, the threshold is dynamically adjusted to absorb them while maintaining low delay. This dynamic adaptation resolves the fixed trade-off between latency and throughput.
Solution Approach 2:
The system performs preliminary detection of oscillation patterns and proactively adjusts thresholds before significant queue buildup occurs. By detecting early signs of oscillation and preemptively modifying the threshold, the system prevents the need for large queuing buffers while maintaining throughput stability, thus achieving both low latency and high throughput.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
According to an embodiment a packet forwarding device is disclosed for forwarding data packets on a link within a communication network. The racket forwarding device is further configured to perform the following steps: measuring a load of the link; detecting if the load exceeds one of a plurality of threshold indicative for a level of congestion on the link; and sending a signal to another device in the communication network signalling the level of congestion.