Island-Based Network Flow Processor for High-Speed Traffic Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Classic network traffic management systems, relying on a single processor, are insufficient for high-speed networks exceeding 100 gigahertz due to processing limitations and queue management inefficiencies.
Innovation Solution
The Island-Based Network Flow Processor (IB-NFP) architecture, featuring multiple interconnected processor islands with a configurable Command/Push/Pull data bus, allows for efficient complex network traffic management by distributing processing tasks across multiple processors and memory systems, optimizing data transmission and queue management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single processor is used for network traffic management, then device complexity is reduced, but processing speed and throughput are insufficient for high-speed networks exceeding 100 gigahertz
Solution Approach 1:
The system divides network traffic management into multiple independent processor islands (first processor island, second processor island, third processor island), each handling specific processing tasks. This segmentation enables parallel processing of network packets, achieving throughput exceeding 100 gigahertz while maintaining manageable complexity through modular architecture
Solution Approach 2:
The patent transitions from a single-processor linear architecture to a multi-dimensional mesh network topology where processors are arranged in islands connected via command mesh, pull-id mesh, and data meshes. This dimensional expansion enables simultaneous multi-path data flow and parallel processing operations
2Productivity
If multiple processors are used to increase processing power, then throughput increases, but memory access efficiency deteriorates due to non-uniform memory access patterns
Solution Approach 1:
Each processor island is equipped with local memory resources (first memory, second memory, third memory) that provide fast access for that specific processor. Critical status information is stored in locally accessible memory, ensuring that frequently accessed data resides close to the processing unit, thereby minimizing memory access time while maintaining high throughput
Solution Approach 2:
The patent introduces a mesh-based interconnection network (command mesh, pull-id mesh, data meshes) as an intermediary between processors and memory systems. This intermediary enables efficient data routing and reduces contention by providing multiple paths for memory access, thereby improving overall memory access efficiency in the multi-processor system
3Speed
If status information is stored in high-speed memory accessible only to one processor, then access speed is improved, but system scalability and collaboration between processors are limited
Solution Approach 1:
The patent implements a hybrid memory architecture where each processor has dedicated high-speed memory for immediate access, while also providing access to shared memory resources through the mesh network. This multi-functional memory system enables both fast local access and collaborative shared access, supporting both speed requirements and processor collaboration needs
Solution Approach 2:
The system merges locally-dedicated memory (first memory, second memory, third memory) with shared memory resources (fourth memory, fifth memory) into a unified memory hierarchy. This combination allows processors to access critical status information locally at high speed while also enabling inter-processor collaboration through shared memory access via the mesh interconnection
Data Source
AI summary
A network appliance includes a first processor, a second processor, a first storage device, and a second storage device. A first status information is stored in the first storage device. The first processor is coupled to the first storage device. A queue of data is stored in the second storage device. The first status information indicates if traffic data stored in the queue of data is permitted to be transmitted. The second processor is coupled to the second storage device. The first processor communicates with the second processor. The traffic data includes packet information. The first storage device is a high speed memory only accessible to the first processor. The second storage device is a high capacity memory accessible to multiple processors. The first status information is a permitted bit that indicates if the traffic data within the queue of data is permitted to be transmitted.