TCP Redundancy Transmission for Delay and Overhead
Solution Overview
Problem
The existing TCP delay performance is inadequate for real-time services due to the inefficiencies of timeout-based retransmission mechanisms and Forward Error Correction (FEC) technologies, which lead to increased network overheads from unnecessary redundancy transmission.
Innovation Solution
A data transmission method that determines redundancy transmission time intervals based on service delay requirements, reducing the total time of these intervals to meet the delay requirements while minimizing unnecessary redundancy transmission by arranging intervals from larger to smaller and adjusting according to feedback receipt.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If timeout-based retransmission mechanism is used, then reliability of data transmission is improved, but delay performance deteriorates due to exponentially increasing time intervals
Solution Approach 1:
The patent applies preliminary action by proactively sending redundant data packets before the original data packet is acknowledged as lost. Instead of waiting for timeout to trigger retransmission, the system pre-sends backup packets that can immediately replace lost data, thereby eliminating the exponential delay inherent in timeout-based mechanisms while maintaining transmission reliability
2Loss of time
If Forward Error Correction (FEC) technology is used, then TCP delay performance is improved, but network overhead increases due to unnecessary redundancy transmission determined only by network status
Solution Approach 1:
The patent implements feedback by using ACK packets from the receiver to dynamically control redundant transmission. The sender monitors whether original data packets are successfully received and adjusts subsequent redundant packet transmission accordingly, stopping redundant transmission when original packets are confirmed received. This feedback mechanism eliminates unnecessary redundancy transmission while maintaining improved delay performance
3Reliability
If redundancy transmission is increased to improve reliability, then data loss probability is reduced, but network overhead and unnecessary transmission increase
Solution Approach 1:
The patent applies dynamics by making the quantity of redundant packets adaptive rather than fixed. The system dynamically adjusts the number of redundant packets based on real-time transmission status, sending more redundancy when packet loss is detected and reducing or stopping redundancy when transmission is successful. This dynamic approach optimizes the balance between reliability and network overhead
Data Source
Figure 1
Figure 2
Figure 3
AI summary
This application provides a data transmission method and apparatus. The method includes: determining a redundancy time according to a delay requirement of a target service, where the redundancy time indicates a time used for sending redundant data of original data of the target service, and the redundancy time is less than or equal to a time corresponding to the delay requirement; determining a redundancy quantity N, where the redundancy quantity N indicates a quantity of times of sending the redundant data, and N is a positive integer; determining N redundancy sending time intervals according to the redundancy time and the redundancy quantity, where a time amount of the N redundancy sending time intervals is less than or equal to the redundancy time; and after the original data of the target service is sent, if feedback data is not received within an nth redundancy sending time interval, sending the redundant data of the original data for the nth time according to the nth redundancy sending time interval, where n is 1, ..., N. Therefore, in this application, not only TCP delay performance can be improved, but also unnecessary redundancy sending can be effectively reduced.