Adaptive Duplicate Acknowledgement Threshold for Data Flow Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data transmission protocols face challenges in accurately determining data segment loss, as indirect indications like time-outs and duplicate acknowledgements can be caused by reordering or delays, leading to inefficient retransmission mechanisms.
Innovation Solution
A method to control data segment flow by setting a duplicate acknowledgement threshold based on the receive window size, allowing for adaptive decision-making on retransmitting segments, ensuring timely retransmission before the receive buffer overflows or stalls.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If retransmission is triggered by time-out or duplicate acknowledgements, then data loss can be detected and retransmitted, but false retransmissions occur due to reordering or delays
Solution Approach 1:
The patent dynamically adjusts the duplicate acknowledgement threshold parameter based on the receive window size. When the receive window is large, the threshold is increased to allow more duplicate acknowledgements before triggering retransmission, reducing false retransmissions due to reordering. When the receive window is small, the threshold is decreased to enable faster detection of actual data loss. This parameter adaptation resolves the contradiction by making the retransmission trigger more context-aware.
Solution Approach 2:
The system transitions from a static retransmission threshold to a dynamic one that adapts to current network conditions and buffer states. The threshold is continuously adjusted based on the receive window size, allowing the system to optimize between early retransmission (when buffer is full) and later retransmission (when buffer has space), thereby reducing unnecessary retransmissions while maintaining reliability.
2Reliability
If retransmission occurs early upon receiving duplicate acknowledgements, then buffer overflow is prevented, but unnecessary retransmissions increase due to reordering
Solution Approach 1:
The patent changes the retransmission threshold parameter dynamically based on the receive window size. When the receive window indicates ample buffer space, the threshold is raised to delay retransmission, avoiding unnecessary retransmissions due to reordering. When the buffer is nearly full, the threshold is lowered to trigger earlier retransmission, preventing buffer overflow. This resolves the contradiction by adapting the retransmission timing to current buffer conditions.
Solution Approach 2:
The system uses feedback from the receive window size to adjust the retransmission threshold. The receiver's buffer status is communicated back to the sender through the window size parameter, which then informs the duplicate acknowledgement threshold setting. This feedback loop enables the sender to make informed retransmission decisions that balance buffer protection with transmission efficiency.
3Reliability
If the duplicate acknowledgement threshold is set low, then retransmission occurs quickly preventing buffer overflow, but network stalls occur due to premature retransmission
Solution Approach 1:
The patent dynamically adjusts the duplicate acknowledgement threshold based on the receive window size to prevent both buffer overflow and network stalls. When the window is large, the threshold is increased to allow more data to be in flight, reducing premature retransmissions and network stalls. When the window is small, the threshold is decreased to prevent buffer overflow. This adaptive parameter adjustment resolves the contradiction by balancing early detection with efficient data flow.
Data Source
AI summary
A method of controlling the flow of an amount of data from a sending peer to a receiving peer of a predetermined communication protocol is described. The method comprises dividing the amount of data into a plurality of data segments, where the data segments are ordered in a sequence. The segments are sent to the receiving peer in the order of said sequence. The receiving peer acknowledges the correct receipt of a data segment and identifies the last correctly received data segment of the sequence that was received in the proper order of the sequence. The sending peer is arranged such that if it receives a threshold number of duplicate acknowledgements, it performs a retransmission. The threshold number that trigger a retransmission is an adaptive parameter determined based on the size of a receive window, the size of the receive window representing a number of data segments which can be stored in a buffer space at the receiving peer.