Switching Network Traffic Control via Queue Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic control methods in switching networks are ineffective in managing congestion in unicast and multicast communication modes, leading to unreliable and inaccurate traffic control responses and a high risk of cache overflow due to low counter-pressure speed.
Innovation Solution
A method and device that monitor the lengths of unicast and multicast virtual output queues in real-time, generate congestion indication signals, and add link-level traffic control signals to cell headers to manage congestion, ensuring accurate and reliable traffic control by distinguishing between unicast and multicast traffic and preventing cache overflow.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If N*N Virtual Input Queues are created to track congestion source, then congestion positioning capability is improved, but device complexity and resource consumption increase hugely
Solution Approach 1:
The patent extracts the congestion positioning function from the queue structure itself by using separate congestion detection units that independently monitor each output queue. This allows the system to identify congested queues without requiring a complex N*N virtual input queue structure, thereby reducing device complexity while maintaining congestion positioning capability.
Solution Approach 2:
The patent introduces congestion detection units as intermediary components that mediate between the output queues and the control logic. These units provide congestion status information without requiring the switching network to directly analyze control cells or maintain complex N*N queue structures, simplifying the overall system architecture.
2Productivity
If input bandwidth is matched with output bandwidth, then bandwidth utilization is improved, but traffic control accuracy deteriorates due to dynamic traffic intensity changes
Solution Approach 1:
The patent implements dynamic bandwidth allocation by separately controlling unicast and multicast traffic based on their respective congestion states. Instead of using a fixed input-output bandwidth match, the system dynamically adjusts bandwidth allocation according to real-time congestion detection results, allowing accurate traffic control despite dynamic traffic intensity changes.
Solution Approach 2:
The patent segments the traffic control mechanism into separate unicast and multicast control paths. This segmentation allows independent bandwidth management for each traffic type, improving both bandwidth utilization and control accuracy by tailoring the bandwidth allocation strategy to the specific characteristics of each traffic type rather than using a single matched bandwidth approach.
3Device complexity
If unicast and multicast traffic control are combined, then system simplicity is improved, but traffic control reliability deteriorates due to mutual interference
Solution Approach 1:
The patent applies segmentation by creating separate control mechanisms for unicast and multicast traffic. Each traffic type has its own congestion detection and control logic, preventing mutual interference while maintaining manageable system complexity. This segmented approach ensures that control actions for one traffic type do not adversely affect the other.
Solution Approach 2:
The patent implements local quality by applying different control strategies tailored to the specific requirements of unicast and multicast traffic. Unicast traffic receives control based on destination-specific congestion states, while multicast traffic receives control based on group-specific congestion states. This localized control approach improves reliability by addressing the unique characteristics of each traffic type without requiring a completely separate system for each.
4Reliability
If counter-pressure speed is increased to prevent cache overflow, then cache overflow risk is reduced, but system stability may deteriorate due to aggressive traffic control
Solution Approach 1:
The patent applies preliminary anti-action by implementing congestion detection and control before cache overflow occurs. The system continuously monitors output queue lengths and proactively adjusts input traffic rates when congestion is detected, preventing cache overflow rather than reacting after it occurs. This preventive approach reduces overflow risk while maintaining stable traffic flow through gradual adjustments.
Solution Approach 2:
The patent implements feedback mechanisms where congestion detection units continuously monitor queue states and provide information to control logic that adjusts traffic rates accordingly. This closed-loop feedback system enables the network to respond to congestion conditions and prevent cache overflow while maintaining stability through controlled, incremental adjustments rather than aggressive traffic control.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
The present disclosure relates to a method and device for controlling the switching network traffic. The method comprises the following steps: monitoring the length of a unicast virtual output queue and a multicast virtual output queue in real time, and when the length of one of the said queues exceeds a congestion threshold, generating a congestion indication signal and adding a congestion identifier to the header of every cell of the corresponding output port (S101); monitoring in real time whether there is any cell entering/leaving the congestion queue, wherein if it does, transmitting a notification message of cell entering/leaving queue (S102); determining whether the number of the cells entering the congestion queue reaches a pre-determined threshold according to the notification message of cell entering/leaving queue, wherein if it does, adding an indication signal of effective link level traffic control to the header of every cell output by the corresponding output port (S103). The device comprises a congestion-monitoring module and a counter-pressure module. The present disclosure solves the problem that the existing solution method is unreasonable in solving the problem of congestion in unicast/multicast modes.