Switch Fabric Hybrid Load Balancing via Capacity Indicators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data transmission from a source device to a destination device via a physical link can be disrupted due to congestion at queues, leading to head-of-line blocking and buffer overflow, with some queues being congested while others are underutilized, necessitating methods to direct data based on destination limitations.
Innovation Solution
A switch module that receives an order identifier and available capacity indicators for multiple modules in a multi-stage switch fabric, directing data packets to modules with higher available capacity to minimize congestion, and routing around congested paths based on load balancing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data packets are directed to queues based on destination limitations, then data transmission reliability is improved, but queue congestion and head-of-line blocking worsen
Solution Approach 1:
The patent implements dynamic load balancing by continuously monitoring queue capacity indicators and adaptively selecting output queues based on current congestion levels. The system transitions from static destination-based routing to dynamic routing that considers real-time queue status, allowing the switch fabric to respond to changing traffic patterns and prevent congestion before it occurs.
Solution Approach 2:
The system changes the routing parameter from solely destination-based to a hybrid approach that incorporates queue capacity indicators. By introducing capacity-based selection as an additional parameter, the system can differentiate between congested and available queues destined for the same output, thereby improving both reliability and preventing congestion.
2Speed
If data packets are routed through congested queues, then transmission speed is maintained, but data loss increases due to buffer overflow
Solution Approach 1:
The system performs preliminary assessment of queue capacity indicators before routing data packets. By checking the available capacity of output queues in advance and selecting queues with sufficient buffer space, the system prevents data loss due to buffer overflow while maintaining transmission speed through efficient load distribution.
Solution Approach 2:
The patent implements a feedback mechanism where queue capacity indicators are continuously monitored and used to influence routing decisions. This closed-loop control allows the system to respond to congestion conditions in real-time, redirecting traffic away from overloaded queues to prevent buffer overflow and data loss.
3Productivity
If hybrid load balancing is implemented, then congestion is minimized, but switch fabric complexity increases
Solution Approach 1:
The patent segments the load balancing function into distinct components: queue capacity indicator generation, indicator monitoring, and routing decision-making. This modular approach allows the hybrid load balancing to be implemented through coordinated actions of multiple switch modules rather than a monolithic complex system, making the complexity more manageable and maintainable.
Data Source
AI summary
In some embodiments, an apparatus includes a switch module configured to receive an order identifier of a first data packet from a first stage of a multi-stage switch. The switch module is configured to receive an indicator of an available capacity of the first module of a second stage of the multi-stage switch fabric, and an indicator of an available capacity of a second module of the second stage of the multi-stage switch fabric. The switch module is configured, when the order identifier is assigned, to direct the first data packet to the first module of a second stage of the multi-stage switch fabric when the available capacity of the second module is lower than the available capacity of the first module. The switch module configured, when the order identifier is unassigned, to direct the first data packet to the second module when the available capacity of the second module is higher than the available capacity of the first module.