Forward Packet Recovery Using Byte-Based Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing forward packet recovery methods in network communications rely on transmitting parity packets based on the number of data packets, leading to unpredictable network bandwidth usage due to varying packet sizes, which can result in inefficient use of network resources and performance issues.
Innovation Solution
The method involves dynamically generating error correcting frames by grouping, fragmenting, or segmenting data packets to achieve a target byte protection ratio, ensuring that error correction information is transmitted based on the number of bytes rather than packets, thereby optimizing network overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If parity packets are transmitted based on the number of data packets, then forward packet recovery capability is provided, but network overhead becomes unpredictable and bandwidth utilization deteriorates
Solution Approach 1:
The patent changes the basis for generating parity packets from packet count to byte count. Specifically, it uses a byte protection ratio (e.g., 1 byte of parity per 4 bytes of data) rather than a packet protection ratio, which makes the amount of parity information predictable and proportional to actual data volume transmitted, thereby improving bandwidth utilization while maintaining recovery capability
Solution Approach 2:
The system dynamically adjusts the number of parity packets generated based on the actual byte size of data packets. When data packets are larger, more parity bytes are generated proportionally; when smaller, fewer parity bytes are generated. This dynamic adaptation ensures consistent bandwidth overhead regardless of packet size variations
2Reliability
If error correcting information is generated based on packet count, then packet loss recovery is enabled, but network overhead increases in constrained network environments
Solution Approach 1:
The invention transitions from using packet count as the basis for error correction to using byte count. A target byte protection ratio is established (e.g., 20% overhead), and parity bytes are generated to match this ratio rather than generating fixed parity packets per data packet, thereby reducing overhead in constrained networks
Solution Approach 2:
Instead of providing full packet-level redundancy for every data packet, the system provides partial redundancy at the byte level proportional to the actual data volume. This partial action approach reduces the total amount of parity information needed while still enabling recovery of lost data
Data Source
AI summary
Disclosed herein are systems and methods for forward packet recovery in a communication network with constrained network bandwidth overhead. In exemplary embodiments, a target byte protection ratio is determined. Error correcting frames are dynamically generated by a first processor such that error correcting information can be generated to approximate the target byte protection ratio. The data packets and error correcting information are then transmitted across one or more communication networks to a second processor. The second processor can use the error correcting information to regenerate or replace data packets missing or corrupted in transmission across one or more communication networks.