Dynamic Queue Priority Reallocation for Network Overload
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network switches, routers, and gateways face issues with instantaneous network overload, leading to frame drops due to unused queues in low activity ports while other ports are overstressed, degrading Quality of Service (QoS).
Innovation Solution
Implementing a controller that dynamically adjusts Internal Priority Values (IPV) based on queue utilization, triggering events like Queue Nearly Full Alert (QNFA) to re-orient high priority frames to low priority queues with available buffer space, optimizing memory usage and preventing frame drops.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If fixed depth queues and queue management algorithms per queue are implemented, then queue management is simplified, but frame drops occur during network overload when some queues are unused while others are collapsed
Solution Approach 1:
The patent implements dynamic queue depth allocation where the controller continuously monitors queue utilization and adjusts the depth of individual queues in real-time. When a queue approaches full capacity, the controller dynamically reduces its depth and redirects subsequent frames to alternative queues, preventing frame drops while maintaining simple per-queue management algorithms.
Solution Approach 2:
The system employs feedback mechanisms where the controller monitors the utilization state of each queue and uses this information to dynamically adjust queue depths. The controller receives feedback about queue fullness and responds by reallocating buffer space, creating a closed-loop control system that adapts to changing traffic conditions and prevents overload-induced frame drops.
2Reliability
If queues of low activity ports are maintained with fixed depth, then memory is allocated for potential future traffic, but this memory remains unused while queues of overstressed ports collapse
Solution Approach 1:
The patent transforms static queue depth allocation into a dynamic system where queue depths are continuously adjusted based on real-time utilization monitoring. Queues for low-activity ports can expand when needed, while queues for currently overloaded ports are restricted, optimizing memory utilization efficiency without sacrificing capacity availability.
Solution Approach 2:
The system changes the parameter of queue depth from a fixed value to a dynamically adjustable parameter. The controller modifies the depth parameter of individual queues based on their utilization state, allowing memory resources to be reallocated from underutilized queues to overloaded queues, thereby improving both capacity availability and memory efficiency.
3Reliability
If internal priority values are dynamically adjusted based on queue status, then frame drops are prevented by reallocating buffer space, but controller complexity increases
Solution Approach 1:
The patent implements a self-service mechanism where the system automatically monitors queue utilization and performs dynamic priority adjustments without external intervention. The controller autonomously detects queue fullness and reallocates buffer space by modifying internal priority values, reducing the need for complex external control logic while maintaining high frame delivery reliability.
Solution Approach 2:
The dynamic IPV adjustment mechanism relies on feedback from queue utilization monitoring. The controller receives feedback about queue states and automatically responds by adjusting internal priority values, creating a self-regulating system that prevents frame drops through real-time adaptation without requiring complex manual configuration or external control.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure relates to controlling queue release in a network. In particular, the disclosure proposes a controller being configured to: obtain a state of each of a plurality of queues of a network node; determine, based on the states of the queues, whether the utilization of one or more queues exceeds one or more thresholds; generate one or more new priority entries for one or more queues of the plurality of queues, if one or more thresholds are exceeded; and provide the one or more new priority entries to the one or more queues of the network node. Further, the disclosure proposes a network node being configured to provide a state of each of a plurality of queues to a controller, and obtain one or more new priority entries for one or more queues of the plurality of queues from the controller.