Dynamic Switch Queue Sizing for Predicted Network Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing buffer management techniques in data centers fail to effectively allocate buffer space dynamically, leading to packet loss due to changing network conditions and varying application requirements, resulting in degraded performance and inefficient use of bandwidth.
Innovation Solution
A method for managing traffic in communication networks that involves training engines to classify traffic units, predicting future traffic volume, and dynamically allocating queue sizes based on predicted loads, using a formula that considers current and future buffer occupancy, congested queues, and dequeue rates to optimize buffer allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If buffer size is increased to accommodate more traffic, then packet loss is reduced, but device complexity and resource utilization efficiency deteriorate
Solution Approach 1:
The patent implements dynamic buffer size adjustment by training machine learning engines to predict future traffic volume and automatically allocate buffer resources accordingly. The buffer size is no longer static but adapts continuously based on predicted traffic patterns, resolving the contradiction between having sufficient buffer capacity and maintaining simple device management.
Solution Approach 2:
The system employs feedback mechanisms where the performance of the communication network is continuously monitored, and the machine learning engines are retrained based on actual performance outcomes. This closed-loop feedback allows the system to learn from past decisions and optimize buffer allocation dynamically, reducing packet loss while avoiding unnecessary complexity.
2Productivity
If buffer size is dynamically adjusted based on traffic conditions, then network performance is improved, but device complexity increases
Solution Approach 1:
The patent enables the buffer management system to serve itself by implementing automated machine learning models that independently predict traffic patterns and adjust buffer allocation without manual intervention. The system self-optimizes by continuously learning from network performance data, improving throughput while keeping operational complexity manageable through automation.
Solution Approach 2:
The system changes the parameter of buffer size dynamically based on predicted traffic conditions rather than using fixed allocations. By adjusting this critical parameter adaptively, the network achieves improved throughput and resource utilization while the complexity is managed through algorithmic control rather than manual configuration.
3Adaptability or versatility
If traditional static buffer allocation is used, then device complexity is reduced, but adaptability to varying traffic patterns deteriorates
Solution Approach 1:
The patent applies preliminary action by training machine learning engines to predict future traffic volume before actual traffic arrives. This advance prediction allows the system to proactively allocate buffer resources according to anticipated demand, significantly improving adaptability to varying traffic patterns while managing complexity through automated forecasting rather than reactive adjustments.
Data Source
AI summary
A method for managing traffic in a communication network and computer program product, the method comprising: receiving plurality of traffic units to be transmitted by a switch through a port having an associated queue; extracting features from the traffic units; providing the features to a first engine, to obtain a class for the traffic units; using a second engine associated with a traffic model for the class, obtaining an indication of a predicted traffic volume for the class for a future time and for the physical location of a switch that transmits the plurality of traffic units; allocating a queue of a size corresponding to the indication; assigning the plurality of traffic units to the buffer; determining whether performance of the communication network has improved or deteriorated; and subject to deterioration of the performance: generating a penalty; and retraining the first engine or the second engine in accordance with the penalty.