FOU Tunnel Load Balancer for Congestion Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network latency and throughput are affected by factors such as the number of hops, congestion, and branching in network paths, leading to congestion hotspots and high latency, which existing technologies fail to effectively address.
Innovation Solution
A method is introduced to select tunnels for network traffic by determining performance metrics, assigning scores, and using a hierarchical tree structure to distribute traffic across multiple tunnels, ensuring fair distribution and avoiding congestion hotspots, utilizing Foo-over-UDP (FOU) tunnels and load balancers to dynamically route traffic based on measured metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If state-less branching based on hash of packet fields is used for traffic distribution, then traffic can be distributed across multiple network paths, but congestion hotspots and high latency occur in some branches
Solution Approach 1:
The patent implements dynamic traffic distribution by transitioning from static hash-based branching to a system that continuously monitors network conditions (latency, throughput, congestion) and adjusts traffic routing in real-time. The load balancer dynamically selects optimal paths based on current network state, resolving the contradiction between maintaining traffic distribution and avoiding congestion hotspots.
Solution Approach 2:
The system incorporates feedback mechanisms where network performance metrics are continuously measured and fed back to the load balancer. This feedback loop enables the system to identify congestion hotspots and adjust traffic distribution accordingly, ensuring that traffic is routed through optimal paths while maintaining overall distribution efficiency.
2Productivity
If multiple tunnels are used for network traffic, then throughput can be increased, but determining optimal tunnel selection becomes complex
Solution Approach 1:
The patent introduces a load balancer as an intermediary component that manages tunnel selection. Instead of distributing the complex selection logic across multiple devices, the load balancer centralizes the decision-making process by evaluating performance metrics of multiple tunnels and selecting optimal ones, thereby increasing throughput while managing selection complexity in a single location.
Solution Approach 2:
The system implements self-service mechanisms where the load balancer automatically monitors tunnel performance and makes routing decisions without manual intervention. Performance metrics are automatically collected, analyzed, and used to adjust traffic distribution, enabling the system to adapt to changing network conditions while maintaining simplicity in operation.
Data Source
AI summary
An approach is disclosed for selecting tunnels through which network traffic can be sent to steer the traffic away from congestion hot-spots. In one embodiment, multiple Foo-over-UDP (FOU) tunnels, each having a distinct source port, are created between two endpoints. Probes are scheduled to measure network metrics, such as latency and liveliness, of each of the FOU tunnels. In turn, the network metrics are used to select particular FOU tunnel(s) to send traffic over in a manner that is fair across source and destination IP addresses of the FOU tunnel(s). In particular, scores are assigned to the source and destination IP addresses based on sums of scores assigned to tunnels having those source and destination IP addresses based on the tunnels' performance metrics. A load balancer then splits a number of buckets across the source and destination IP addresses, and ultimately across the tunnels, based on the assigned scores.