Bonding Engine Dynamic Packet Fragmentation for Latency Minimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing link bonding systems face issues with delay and jitter sensitivity, leading to undesirable throughput and latency performance, particularly in upstream data rates over EFM-bonding ADSL loops, and fragmentation algorithms do not effectively minimize delay across bonded links.
Innovation Solution
A communications system with a bonding engine that fragments packets based on a cost function considering differential delay, payload rate, and buffered data across multiple subscriber line pairs, dynamically adapting fragment sizes to minimize latency and equalize cost functions across the bonding group.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If packets are fragmented across multiple subscriber line pairs in a bonding group, then throughput is improved, but latency and delay variations increase
Solution Approach 1:
The system dynamically adjusts packet fragment sizes based on real-time conditions of each subscriber line pair, including differential delay, payload rate, and buffered data. This dynamic adaptation allows the system to optimize the balance between throughput and latency for each packet transmission, rather than using a static fragmentation approach.
Solution Approach 2:
The invention changes the parameters of packet fragmentation by calculating optimal fragment sizes based on cost functions that consider differential delay, payload rate, and buffered data. By varying these parameters per packet rather than using fixed fragmentation, the system achieves better throughput-latency tradeoff.
2Device complexity
If fixed fragmentation algorithms are used across bonding groups, then implementation is simplified, but delay and jitter sensitivity increases
Solution Approach 1:
Instead of using fixed fragmentation parameters, the system calculates optimal fragment sizes dynamically based on real-time measurements of differential delay, payload rate, and buffered data for each subscriber line pair. This adaptive approach reduces delay and jitter sensitivity while remaining implementable through standardized processing steps.
Solution Approach 2:
The system incorporates feedback mechanisms by continuously monitoring the actual transmission characteristics of each subscriber line pair and using this information to adjust fragmentation parameters. The cost function uses measured differential delay and buffered data to optimize subsequent packet fragmentation decisions.
Data Source
AI summary
A telecommunications access module comprises at least one bonding engine that receives packet data and bonds subscriber line pairs together to form a bonding group over which packet fragments are transmitted. A database maintains a list of communications subscriber line pairs that form the bonding group. A processor is configured to determine whether an incoming packet subject to fragmentation is long enough to fragment across multiple subscriber line pairs. If the incoming packet can be fragmented, the processor is configured to calculate a cost function for minimum packet fragment sizes that is proportional to the end-of-fragment arrival time at the far end of each subscriber line pair and calculate remainders of packet fragment length. The packet is fragmented based on the calculated cost function and remainders of any packet fragment lengths and allocated to the subscriber line pairs and transmitted.