Per-Flow Switch Architecture for Fast Network Congestion Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network architectures face challenges in scalability, versatility, and efficiency due to increasing network load and diverse traffic types, with conventional congestion control methods being slow and ineffective, leading to suboptimal network utilization.
Innovation Solution
A switch architecture that maintains state information of individual packet flows, using flow-specific input queues and acknowledgments to enable dynamic flow control, allowing each switch to manage traffic on a per-flow basis, thus accommodating dynamic traffic with fast and effective congestion control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional network architectures are used, then network load handling is basic, but network utilization is suboptimal and congestion control is slow
Solution Approach 1:
The patent segments network traffic into individual packet flows, with each flow receiving dedicated state information tracking and flow-specific input queues. This segmentation enables independent congestion control for each flow, improving overall network utilization while providing rapid, targeted congestion responses rather than blanket network-wide control.
Solution Approach 2:
The patent implements feedback mechanisms where switches maintain state information about each packet flow and use this information to dynamically adjust flow control decisions. The system continuously monitors flow states and provides feedback to control congestion in real-time, enabling fast response to congestion conditions while optimizing network utilization.
2Measurement precision
If per-flow state information is maintained, then flow control precision is improved, but switch device complexity increases
Solution Approach 1:
The patent divides the switch architecture into modular components, with each input port having dedicated flow state memory and flow-specific queues. This modular segmentation allows precise per-flow control while distributing complexity across multiple independent units rather than requiring a monolithic complex system.
Solution Approach 2:
The patent adds a new dimension of flow state information storage at the input port level, creating a multi-dimensional data structure that tracks flow states independently for each input port and flow combination. This dimensional expansion enables precise flow control without requiring complex cross-port state management.
3Adaptability or versatility
If dynamic flow control is implemented, then network adaptability to diverse traffic is improved, but control mechanism complexity increases
Solution Approach 1:
The patent implements dynamic flow control where state information is created, modified, and removed based on real-time flow conditions. Flows are dynamically established when packets arrive and dynamically removed when flows complete, allowing the system to adapt to diverse traffic patterns while maintaining a relatively simple underlying architecture through automated dynamic behavior.
Solution Approach 2:
The patent enables flows to self-manage their own control state through automated processes. When packets arrive, the system automatically creates flow state information; when packets are transmitted, the system automatically updates state; when flows complete, the system automatically removes state. This self-service approach provides high adaptability without requiring complex external control mechanisms.
Data Source
AI summary
A switch architecture for a data-driven intelligent networking system is provided. The system can accommodate dynamic traffic with fast, effective congestion control. The system can maintain state information of individual packet flows, which can be set up or released dynamically based on injected data. Each flow can be provided with a flow-specific input queue upon arriving at a switch. Packets of a respective flow are acknowledged after reaching the egress point of the network, and the acknowledgement packets are sent back to the ingress point of the flow along the same data path. As a result, each switch can obtain state information of each flow and perform flow control on a per-flow basis.