Unified Packet Recirculation Port for Bandwidth and Latency Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing packet recirculation architectures in network devices are non-unified, leading to increased complexity, requiring separate ports and programming models for loopback and egress recirculation, which limits support for additional features and introduces latency and bandwidth constraints.
Innovation Solution
A unified packet recirculation approach that integrates ingress and egress pipelines, eliminating the need for separate ports and using a consistent programming model, allowing simultaneous multiple copies and different classes of service without out-of-order processing, thereby optimizing bandwidth and reducing complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If non-unified packet recirculation architecture is used, then separate ports and programming models can be implemented, but device complexity increases and bandwidth is constrained
Solution Approach 1:
The patent merges separate loopback and egress recirculation ports into a single unified recirculation port. This consolidation eliminates the need for distinct hardware ports and separate programming models, thereby reducing device complexity while maintaining recirculation functionality. The unified port handles both loopback and egress recirculation traffic through a common interface and control mechanism.
Solution Approach 2:
The unified recirculation port is designed to perform multiple functions: it handles both loopback recirculation and egress recirculation operations. This multi-functional design eliminates the need for separate dedicated ports, reducing overall system complexity while preserving the adaptability and versatility of the recirculation architecture.
2Reliability
If non-unified packet recirculation architecture is used, then separate processing paths can be established, but bandwidth is limited and latency increases
Solution Approach 1:
By merging the recirculation paths into a unified architecture, the patent enables more efficient utilization of available bandwidth. The unified port allows packets to be recirculated without the overhead of separate port handling and reduces the latency associated with multiple processing paths, thereby improving productivity while maintaining reliability through consistent packet handling.
3Ease of operation
If separate ports are used for loopback and egress recirculation, then dedicated processing can be provided, but hardware overhead increases
Solution Approach 1:
The patent combines the functionality of separate loopback and egress recirculation ports into a single unified recirculation port. This merger reduces hardware overhead by eliminating redundant port structures and associated control logic, while the unified programming model maintains ease of operation by providing a consistent interface for both recirculation types.
4Productivity
If unified packet recirculation is implemented, then bandwidth is improved to 13.5 Gbps, but new programming challenges arise
Solution Approach 1:
The unified recirculation port employs a universal programming model that handles both loopback and egress recirculation through a single interface. This multi-functional programming approach simplifies the software layer by eliminating the need for separate programming logic for different recirculation types, thereby managing programming complexity while enabling high bandwidth performance of 13.5 Gbps.
Data Source
AI summary
An apparatus includes an ingress data buffer, an ingress processor, an egress processor, and a recirculation data buffer. The apparatus is configured to provide unified packet recirculation via the recirculation data buffer and a single recirculation port on the ingress data buffer. The apparatus can be a switch, router or other network device.