Network Device Training Plane Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network devices struggle with complete automation due to the lack of support for generating control commands using internal algorithms, limiting their ability to perform diverse control operations efficiently.
Innovation Solution
Introducing a training plane function that utilizes a machine learning model to analyze network data and generate control commands, separating the network device into control plane and training plane functions for enhanced automation and flexible control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If network devices use internal algorithms for control operations, then they can perform control functions, but they cannot support complete automation
Solution Approach 1:
The network device is divided into two distinct planes: a training plane that handles machine learning model training and generation of control commands, and a control plane that executes these commands. This segmentation allows the device to achieve complete automation through the training plane while maintaining versatile control operations through the control plane, resolving the contradiction between automation capability and control flexibility.
2Extent of automation
If network devices separate control plane and training plane functions, then complete automation is supported, but device complexity increases
Solution Approach 1:
By segmenting the network device into control plane and training plane functions, the patent achieves complete automation while managing complexity through clear functional separation. The training plane handles AI model training and command generation, while the control plane handles command execution, allowing each component to be optimized independently.
Solution Approach 2:
The patent introduces a model management server as an intermediary that assists in managing machine learning models. This external intermediary handles complex model training and updates, reducing the burden on the network device itself while still enabling complete automation capability.
3Productivity
If machine learning models are used to generate control commands, then operational efficiency improves, but implementation cost increases
Solution Approach 1:
The model management server acts as an intermediary that handles the complex and expensive task of machine learning model training and management. By outsourcing this function to a specialized external server, the patent can leverage ML-based control command generation for improved operational efficiency while avoiding the high implementation costs of building and maintaining ML infrastructure within each network device.
Solution Approach 2:
The model management server provides a universal platform that can serve multiple network devices with the same machine learning models. This multi-functionality allows the system to achieve operational efficiency improvements across the network while amortizing the implementation cost across multiple devices, making the solution more economically viable.
Data Source
AI summary
A network device divided into a training plane and a control plane, model management server that controls a network device, and processing methods of a network device and model management server are disclosed. A processing method may include receiving a machine learning model from a model management server, obtaining network data to generate analytics information, generating analytics information by inputting the network data to a machine learning model, feeding back the analytics information to the model management server, and generating a control command of the network device using the analytics information, wherein the analytics information is generated by a training plane function and the control command is generated by a control plane function.