Flow Rate Network Load Balancing Spine-Leaf Congestion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In spine-leaf networks, existing load balancing techniques face challenges in efficiently identifying and mitigating network-level congestion without introducing significant modifications or relying on specific feedback mechanisms, particularly in scenarios where multiple paths have equal costs, leading to suboptimal traffic distribution.
Innovation Solution
The implementation of flow rate-based load balancing methods, where an ingress top-of-rack switch determines the instantaneous flow rate of a packet flow and compares it to an average flow rate to dynamically assign the packet to either the current path or an alternative path, using locally measured metrics and minimizing network-level modifications, without relying on dedicated feedback mechanisms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional load balancing techniques are used in spine-leaf networks, then traffic is distributed across multiple equal-cost paths, but network-level congestion cannot be effectively identified and mitigated without significant modifications or dedicated feedback mechanisms
Solution Approach 1:
The network system uses its own existing flow rate measurement capabilities to identify congestion, eliminating the need for external feedback mechanisms. The load balancer monitors flow rates on each path using locally available information and automatically reroutes flows when congestion is detected, making the system self-diagnosing and self-healing without requiring additional network infrastructure or protocols
Solution Approach 2:
The patent implements an internal feedback loop where the load balancer continuously monitors flow rates on multiple paths and uses this information to dynamically adjust flow routing decisions. When the flow rate on a primary path exceeds a threshold indicating congestion, the system feedback-triggers rerouting of affected flows to alternative paths, creating a closed-loop control system that automatically responds to network conditions
2Productivity
If flow rate-based dynamic rerouting is implemented, then congestion is effectively mitigated and traffic distribution is optimized, but the system requires continuous monitoring and comparison of flow rates which increases processing overhead
Solution Approach 1:
The patent transforms the abstract concept of network congestion into a measurable parameter by continuously monitoring flow rates. It establishes a threshold-based decision mechanism where flows are rerouted when flow rate exceeds a predetermined threshold, converting complex network state assessment into simple parameter comparison that minimizes processing overhead while maintaining effective congestion mitigation
Data Source
AI summary
Techniques for flow rate based load balancing are described. In one embodiment, a method includes receiving a packet associated with a packet flow at an ingress top-of-rack switch in a spine-leaf network. The method includes determining identification information for the packet flow, including a source and a destination. Based on the identification information, the method includes determining an instantaneous flow rate for a first path through the network from the source to the destination. The method also includes comparing the instantaneous flow rate to an average flow rate. Based on the comparison between the instantaneous flow rate to the average flow rate, the method includes assigning the packet flow to one of the first path or a second path.