Flexible Ethernet Data Processing for Switching Circuit Resource Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing flexible Ethernet switching circuits face inefficiencies due to a high speedup factor and excessive resource consumption when designed to handle maximum bandwidths, leading to unnecessary waste of circuit resources.
Innovation Solution
The proposed solution involves using a flexible Ethernet data processing method that converts client service flows into parallel client slot flows, allowing for dynamic configuration of input interfaces to match the fixed bandwidth capacity of each slot, thereby optimizing the switching circuit's capacity based on the physical interface's limitations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the switching circuit is designed according to the maximum bandwidth capacity of each client service flow, then the switching capacity is sufficient to handle peak demands, but the speedup factor becomes excessively large and circuit resources are wasted
Solution Approach 1:
The patent segments the client service flow into multiple time slots, each with a fixed bandwidth capacity. The switching circuit is then designed based on the bandwidth capacity of a single time slot rather than the aggregate maximum bandwidth of all service flows. This segmentation allows the switching circuit to be sized appropriately for actual usage patterns while maintaining sufficient capacity for peak demands in individual slots.
Solution Approach 2:
The patent implements dynamic bandwidth allocation where the total bandwidth of a client service flow is distributed across multiple time slots. The switching circuit dynamically switches between different time slots based on actual traffic demands, allowing the system to adapt to varying bandwidth requirements without needing to provision for the maximum possible aggregate bandwidth continuously.
2Adaptability or versatility
If the switching circuit is designed with high bandwidth capacity to accommodate multiple client service flows, then the switching capacity reaches 500G for five 100G flows, but the actual utilization is limited to 100G per PHY interface, resulting in a speedup factor of 5 and excessive resource consumption
Solution Approach 1:
The patent divides the client service flow into discrete time slots, each with a fixed bandwidth capacity. This segmentation allows the switching circuit to be designed based on the bandwidth capacity of a single time slot rather than the aggregate maximum bandwidth of all service flows. For example, if a client service flow has a bandwidth range of [5G, 100G] and is divided into multiple time slots, each slot has a fixed bandwidth that the switching circuit can handle independently.
Solution Approach 2:
The patent changes the bandwidth parameter from a continuous variable to a discrete set of fixed bandwidth values corresponding to different time slots. Each time slot has a specific fixed bandwidth capacity, and the switching circuit is designed to handle this fixed bandwidth rather than the full range of possible bandwidth values. This parameter transformation simplifies the switching circuit design while maintaining flexibility through time-slot multiplexing.
3Reliability
If the switching circuit is designed according to the maximum bandwidth of client service flows, then sufficient capacity is provided for peak demands, but the speedup factor becomes unnecessarily large, wasting circuit resources
Solution Approach 1:
The patent segments the client service flow into multiple time slots, each with a fixed bandwidth capacity. The switching circuit is then designed based on the bandwidth capacity of a single time slot rather than the aggregate maximum bandwidth of all service flows. This segmentation allows the switching circuit to be sized appropriately for actual usage patterns while maintaining sufficient capacity for peak demands in individual slots.
Solution Approach 2:
The patent ensures continuous operation of the switching circuit by maintaining a steady stream of time slots with fixed bandwidth capacities. Rather than designing for peak aggregate bandwidth that may not be continuously utilized, the system maintains continuous throughput at a lower, more sustainable rate that matches actual usage patterns, thereby improving resource utilization efficiency.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
Embodiments of the present invention disclose a flexible-Ethernet data processing method and a related device. The method includes: acquiring a to-be-switched first client service flow, where the first client service flow is a service flow suitable for transmission on a flexible Ethernet, for example, a service flow that meets a flexible Ethernet transmission standard; performing first rate adaptation from a source clock domain to a target clock domain on the first client service flow, to obtain a second client service flow that matches the target clock domain; and performing serial-to-parallel conversion on the second client service flow in the target clock domain, to obtain a parallel client slot flow. In this way, a proper speedup factor of a switching circuit can be acquired, avoiding waste of circuit resources.