Semi-Static Traffic Engineering via Pre-Computed Path Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current traffic engineering methods, particularly semi-static traffic engineering, face challenges in managing network load and optimizing path selection due to high computational and control signaling loads, especially in scenarios with frequent flow activations and deactivations, such as machine-type communication services, which can lead to inefficient network performance and bottlenecks.
Innovation Solution
The method involves pre-computing path selection configurations for a predetermined number of network load representative scenarios, allowing for look-up and implementation when the prevailing network load matches one of these scenarios, reducing the frequency of traffic engineering triggers and associated computational and control signaling loads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If semi-static traffic engineering is used to optimize path selection, then network performance is improved, but computational load and control signaling load increase significantly
Solution Approach 1:
The patent pre-computes path selection configurations for multiple network load scenarios before they are needed. When a traffic engineering trigger occurs, the system simply looks up the pre-computed configuration matching the current load scenario, avoiding the need for complex real-time computations. This preliminary action significantly reduces the computational load during actual traffic engineering operations while maintaining near-optimal path selection.
2Reliability
If semi-static traffic engineering is used to optimize path selection, then network performance is improved, but control signaling load increases
Solution Approach 1:
The system pre-computes and stores path selection configurations for various network load scenarios before triggers occur. When a trigger happens, the controller simply retrieves the matching pre-computed configuration and sends it to affected nodes, dramatically reducing the control signaling load compared to computing paths in real-time. This approach maintains near-optimal routing while minimizing control plane overhead.
3Adaptability or versatility
If frequent flow activations and deactivations occur, then network adaptability is improved, but frequency of traffic engineering triggers increases
Solution Approach 1:
The patent addresses frequent flow activations by pre-computing path selections for multiple network load scenarios in advance. When flows are activated or deactivated, the system can quickly match the new load state to a pre-computed scenario and retrieve the corresponding configuration, avoiding frequent re-computations. This maintains network adaptability to flow dynamics while reducing the frequency of actual traffic engineering trigger executions.
Solution Approach 2:
The system changes the parameter of path selection from dynamic real-time computation to selection from pre-computed options based on load scenario matching. By categorizing network states into discrete load scenarios with pre-determined optimal paths, the system can adapt to frequent flow changes without triggering frequent re-computations, instead simply selecting from pre-prepared configurations that match current load conditions.
Data Source
AI summary
A method and application function for managing data flows in a network comprising a plurality of nodes. The method comprises generating a number N of network load representatives (NLRs) of an expected loading of the network, identifying a path selection configuration for the data flows for each of the N NLRs and mapping a prevailing network loading to a selected one of the N NLRs. If the selected NLR is different from a currently selected NLR, the method triggers traffic engineering (TE) by implementing the path selection configuration of the selected NLR at nodes affected thereby. The application function may comprise an NLR generator, a path optimizer and an NLR mapper.