Multi-Datapath Traffic Manager for Low-Latency CT and SAF Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network technologies face challenges in managing cut-through (CT) and store-and-forward (SAF) traffic, leading to increased latency and inefficiencies, especially in high-throughput networks, due to the need for delay matching between different packet types and shared buffer access.
Innovation Solution
Implementing separate control data paths for CT and SAF traffic with dedicated processing engines and a dequeue request path merger (DRPM) to manage shared resources, ensuring low latency for CT traffic while maintaining high performance for SAF traffic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If separate control data paths are implemented for CT and SAF traffic, then latency for CT traffic is reduced, but device complexity increases
Solution Approach 1:
The patent segments the control data path into separate dedicated paths for CT traffic and SAF traffic. CT traffic follows a direct path with minimal processing steps, while SAF traffic follows a path with full processing including delay matching. This segmentation allows CT packets to bypass unnecessary processing stages, reducing latency without requiring complete path duplication throughout the entire system.
Solution Approach 2:
The patent introduces a dimensional separation in the control data flow by creating parallel processing dimensions - one for expedited CT traffic and one for standard SAF traffic. This allows the system to handle different traffic types with appropriate processing depths simultaneously, achieving low latency for CT while maintaining proper processing for SAF without exponential complexity growth.
2Loss of time
If dedicated cut-through network path is implemented, then latency is reduced, but adaptability to large or complex network environments decreases
Solution Approach 1:
The patent creates a universal control data path structure that can handle both CT and SAF traffic types within the same network device. The segmented path design allows the same hardware infrastructure to serve multiple traffic types with different latency requirements, making the solution adaptable to various network configurations and traffic patterns without requiring dedicated physical paths for each traffic type.
Solution Approach 2:
The patent implements dynamic path selection within the control data flow, allowing packets to be routed through appropriate processing stages based on their type. CT packets dynamically bypass certain processing stages while SAF packets undergo complete processing, enabling the system to adapt to different traffic requirements in real-time without fixed dedicated paths.
3Adaptability or versatility
If frequent transitions and mixtures of CT and SAF traffic occur, then network versatility is improved, but packet re-ordering and latency issues increase
Solution Approach 1:
The patent segments the control data path into separate dedicated paths for CT traffic and SAF traffic. CT traffic follows a direct path with minimal processing steps, while SAF traffic follows a path with full processing including delay matching. This segmentation allows CT packets to bypass unnecessary processing stages, reducing latency without requiring complete path duplication throughout the system.
Solution Approach 2:
The patent introduces an intermediary mechanism in the form of a dequeue request path merger (DRPM) that coordinates between the separate CT and SAF control paths. This mediator ensures proper synchronization and ordering of packets from both paths, preventing re-ordering issues while maintaining the benefits of separate path processing. The DRPM acts as a coordination point that manages the interaction between expedited and standard traffic flows.
Data Source
AI summary
Techniques as described herein may be implemented to support processing CT and SAF traffic. A common packet data buffer is allocated to store incoming CT and SAF packet data. SAF packet control data are directed onto a control data path with first processing engines, to arrive at a scheduler with a first latency. CT packet control data are directed onto a second control data path to arrive at the scheduler with a second latency less than the first latency after processing in the second control path by second processing engines bypassing a subset of the first processing engines. CT and SAF packet dequeue requests are generated for CT and SAF packets, respectively, using the CT and SAF packet control data and merged into a merged sequence of dequeue requests to retrieve corresponding packet data from the common packet data buffer based on the merged sequence of dequeue requests.