Transport-Layer Retransmission Control for Out-of-Order Packet Loss
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In out-of-order data transfers, the uncertainty of whether a packet is delayed or lost causes high latency due to the receiver delaying retransmission requests, leading to increased tail latency and impacting application performance.
Innovation Solution
A network device detects discarded packets due to physical layer errors and notifies connected nodes, allowing them to adjust transport layer parameters to reduce retransmission latency by reducing the time waited before sending retransmission requests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the receiver delays retransmission requests to distinguish between delayed and lost packets, then duplicate packet transmission is prevented, but retransmission latency increases
Solution Approach 1:
The network device performs preliminary detection of packet loss at the physical layer before packets reach the transport layer. By detecting PCS errors and generating loss indicators in advance, the receiver can immediately adjust retransmission timing without waiting for timeout periods, thus preventing duplicate transmissions while reducing latency.
Solution Approach 2:
The system implements feedback through loss indicators sent from the network device to endpoints. These indicators provide real-time information about physical layer packet losses, enabling receivers to adjust retransmission behavior dynamically based on actual network conditions rather than relying on delayed timeout-based detection.
2Measurement precision
If the receiver waits for a predetermined time before requesting retransmission, then uncertainty between delayed and lost packets is resolved, but application performance deteriorates
Solution Approach 1:
The network device performs preliminary detection of packet loss at the physical layer before packets reach the transport layer. By detecting PCS errors and generating loss indicators in advance, the receiver can immediately adjust retransmission timing without waiting for timeout periods, thus preventing duplicate transmissions while reducing latency.
Solution Approach 2:
The loss indicator acts as an intermediary signal that bridges the physical layer and transport layer. It conveys packet loss information from the network device to endpoints, enabling accurate packet status determination without requiring receivers to wait for timeout periods, thus maintaining both precision and performance.
3Loss of time
If all nodes reduce retransmission time after receiving error notifications, then tail latency is reduced, but false positives may cause unnecessary retransmissions
Solution Approach 1:
The system applies local quality by having individual nodes adjust retransmission parameters based on their specific receipt of loss indicators. Only nodes that receive the loss indicator for a particular packet will reduce their retransmission timeout, while other nodes maintain normal timing. This localized response prevents false positives from causing unnecessary retransmissions across the entire network.
Data Source
AI summary
A device, communication system, and method are provided. In one example, a system for reducing latency in out-of-order data transfers is described that includes circuits to detect a network error; and in response to detecting the network error, transmit a notification to at least one of a plurality of connected nodes that the network error was detected, wherein the notification causes each node that receives the notification to change at least one transport layer parameter.