Optical Switch Configuration and Deflection Routing for Low-Latency DIDCs

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional data center architectures face challenges with latency, scalability, and network management complexity, particularly in handling data-intensive applications, and determining optimal configurations for optical switches and routing paths within Directly Interconnected Data Centers (DIDCs) is a key challenge.

Innovation Solution

A machine learning-based approach is employed to jointly determine the configuration of optical switches and routing paths using a neural network, which learns from traffic matrices to provide deflection routing parameters and optical switch configurations, optimizing the network configuration for DIDCs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If traditional data center architectures are used, then network management is simpler, but latency increases and scalability deteriorates

Engineering Contradiction:
ImprovelatencyVSAvoidnetwork management complexity
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent replaces traditional manual network configuration methods with a machine learning model that automatically determines optimal optical switch configurations and routing paths. The ML model processes traffic matrices and directly outputs configuration parameters, substituting the mechanical/manual network management process with an intelligent automated system that reduces both latency and management complexity

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the operational parameters of optical switches from fixed traditional configurations to dynamically optimized settings determined by the machine learning model. By adjusting optical switch configurations and routing parameters based on learned patterns from traffic matrices, the system achieves lower latency while maintaining manageable complexity through automation

Inventive Principle:
Principle #35Parameter changes

2Speed

If optical switches are configured for high performance, then latency is reduced, but determining optimal configurations becomes more complex

Engineering Contradiction:
Improvenetwork operation speedVSAvoidconfiguration optimization difficulty
Core Design Contradiction:
SpeedVSDifficulty of detecting and measuring

Solution Approach 1:

The machine learning model performs self-service by automatically analyzing traffic matrices and generating optimal optical switch configurations without external intervention. The system learns from historical traffic data and autonomously determines the best configuration parameters, eliminating the need for manual optimization efforts while achieving high network operation speed

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent implements a feedback mechanism where the machine learning model continuously learns from traffic matrices and adjusts optical switch configurations accordingly. The model uses backpropagation to refine its predictions based on performance outcomes, creating a closed-loop system that automatically optimizes configurations for reduced latency without increasing operational difficulty

Inventive Principle:
Principle #23Feedback

3Productivity

If joint determination of routing and switch configuration is implemented, then network efficiency is improved, but computational complexity increases

Engineering Contradiction:
Improvenetwork efficiencyVSAvoidcomputational complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent merges the determination of routing paths and optical switch configurations into a single unified machine learning model. Instead of separately optimizing routing and switch configurations, the model jointly processes traffic matrices and simultaneously outputs both routing parameters and switch configuration settings, improving network efficiency while consolidating computational complexity into one integrated system

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP4708900A1Machine-learning based approach for robust configuration of optically interconnected data centers
Publication Date: 2026.03.11 NOKIA SOLUTIONS & NETWORKS OY
  • EP4708900A1 patent drawingFigure 1A
  • EP4708900A1 patent drawingFigure 1B
  • EP4708900A1 patent drawingFigure 1C

AI summary

In some embodiments, there may be provided a systems, methods, and articles of manufacture that learn, by the machine learning model and based at least on the at least one traffic matrix, a first output indicative of at least one deflection routing parameter and a second output indicative of at least one optical switch configuration, wherein the learning jointly determines the first output indicative of at least one deflection routing parameter and a second output indicative of at least one optical switch configuration; provide the first output indicative of the at least one deflection routing parameter to a network management system to configure at least one aggregation node comprised in a directly interconnected data center; and provide the second output indicative of the at least one optical switch configuration to the network management system to configure at least one optical switch comprised in the directly interconnected data center.