SD-WAN Traffic Conditioning with Selective Forward Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Real-time applications such as streaming and interactive communications in SD-WAN networks face delays due to packet retransmissions, which are not effectively managed by existing error correction techniques that increase traffic and latency.
Innovation Solution
Conditioning traffic through multiple data paths in an SD-WAN by monitoring path quality and link utilization, generating and replicating forward error correction (FEC) packets only when necessary, and sending them on less utilized paths to minimize delay and packet loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If forward error correction (FEC) packets are sent to correct packet loss, then packet loss is reduced, but network latency increases due to additional traffic
Solution Approach 1:
The patent applies partial action by selectively generating FEC packets only when packet loss is detected on a path, rather than continuously sending FEC packets. The system monitors path quality and only activates error correction when needed, reducing unnecessary traffic and latency while maintaining reliability when packet loss occurs
Solution Approach 2:
The patent segments the network traffic into data packets and separate FEC packets, allowing them to be transmitted on different paths. This segmentation enables optimized routing where data packets can use any available path while FEC packets are routed on paths with lower utilization, reducing the impact on latency
2Reliability
If FEC packets are generated and sent on the same path as data packets, then error correction is provided, but bandwidth consumption increases on already congested paths
Solution Approach 1:
The patent transitions from a single-dimension approach (sending all packets on one path) to a multi-dimensional approach by utilizing multiple network paths simultaneously. FEC packets are routed on alternative paths with lower utilization, effectively adding a spatial dimension to packet distribution and reducing congestion on primary paths
Solution Approach 2:
The patent applies local quality by assigning different routing strategies to different packet types based on local path conditions. Data packets are sent on the primary path while FEC packets are sent on alternative paths with better availability, optimizing bandwidth usage based on local network conditions at each path
3Reliability
If retransmission is used to handle lost packets, then data integrity is maintained, but delay increases for real-time applications
Solution Approach 1:
The patent applies preliminary action by proactively generating and sending FEC packets along with data packets before any loss occurs. This allows the receiving end to immediately reconstruct lost packets using the pre-sent FEC information, eliminating the need for time-consuming retransmission requests and responses
Data Source
AI summary
Embodiments herein disclose conditioning traffic through multiple data paths of a Software-Defined Wide Area Network (SD-WAN). Some embodiments include monitoring a path through an SD-WAN to reach a destination node, determining a quality score for packets to the destination node on the path, determining a link utilization for the path, sending a data packet sequence to the destination node on the path, generating a forward error correction (FEC) packet for the data packet sequence, and sending the FEC packet to the destination node on the path in response to the quality score being less than a quality threshold and the link utilization being less than a high utilization threshold.