Adaptive Rate Adaptation Using Queuing Delay Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing accelerated ramp up mechanisms in computer networks face challenges in balancing the rate of packet transmission to avoid poor user experience due to slow ramp up and potential congestion from fast ramp up, leading to queuing delays and impaired performance.
Innovation Solution
A mechanism that determines a queuing delay reporting time and adjusts the packet rate based on this time and a maximum allowable delay to ensure that the queuing delay contributed by any rate increase is bounded, preventing congestion and maintaining network performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the packet rate is increased rapidly to improve user experience, then the bandwidth utilization increases, but network congestion occurs leading to queuing delays and performance degradation
Solution Approach 1:
The patent applies dynamics by making the packet rate adjustment adaptive rather than fixed. The rate multiplication factor is dynamically adjusted based on real-time queuing delay measurements, allowing the system to optimize bandwidth utilization while preventing congestion. The rate increases exponentially when delays are low but slows down when delays approach thresholds, creating a dynamic balance between productivity and reliability.
Solution Approach 2:
The patent implements feedback mechanisms by continuously monitoring queuing delay and using this information to adjust the packet transmission rate. The sender measures actual queuing delay experienced by packets and feeds this back into the rate adaptation algorithm, creating a closed-loop system that automatically adjusts rate based on network conditions to prevent congestion while maximizing bandwidth usage.
2Reliability
If the packet rate is increased slowly to avoid network congestion, then network stability is maintained, but user experience deteriorates due to poor video quality and slow session establishment
Solution Approach 1:
The system dynamically adjusts the rate multiplication factor based on real-time network conditions. During initial phase, when queuing delay is low, the rate increases rapidly (exponential growth) to establish sessions quickly. As the rate increases and queuing delay approaches thresholds, the growth rate automatically slows down, creating a dynamic curve that balances fast establishment with stability.
Solution Approach 2:
The patent applies preliminary action by implementing a ramp-up phase during initial session establishment where the packet rate is deliberately increased faster than normal operation. This preliminary rapid deployment establishes the session quickly before transitioning to more conservative rate increases, ensuring fast startup without causing permanent congestion issues.
3Productivity
If the packet rate is increased aggressively to quickly use excess capacity, then bandwidth efficiency improves, but queuing delay increases leading to packet losses and impaired end-to-end performance
Solution Approach 1:
The system continuously measures actual queuing delay experienced by packets and uses this feedback to control the rate multiplication factor. When queuing delay remains below thresholds indicating available capacity, the rate can be multiplied aggressively (e.g., by factor of 2 or more) to maximize bandwidth efficiency. When delay approaches critical thresholds, the feedback mechanism reduces the multiplication factor to prevent excessive queuing and packet losses.
Solution Approach 2:
The patent changes the parameter of rate multiplication factor based on network conditions. The multiplication factor is not fixed but varies dynamically - using higher factors (e.g., 2x, 4x) when capacity is available and lower factors when approaching congestion thresholds. This parameter adaptation allows aggressive rate increases for efficiency while automatically backing off when queuing delay becomes problematic.
Data Source
AI summary
In one embodiment, a device in a network sends a first set of one or more packets at a first rate to a destination in the network. The device determines a queuing delay reporting time. The queuing delay reporting time represents an amount of time before an indication of a queuing delay associated with the sent first set of one or more packets is reported back to the device. The device determines a second rate based on the determined queuing delay reporting time and on a maximum allowable delay. The device sends a second set of packets at the second rate to the destination.