Computing Node Resource Configuration for Model Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Updating machine learning or computational models on constrained devices requires full firmware updates, leading to large data transfers and operational issues, as traditional firmware updates are generalized across devices rather than tailored to specific devices, resulting in inefficiencies and potential misconfiguration.
Innovation Solution
A method for a computing node to receive and configure configuration information for resources associated with computational operations, allowing for the execution of these operations without the need for full firmware updates, by exposing and managing resources via Lightweight Machine to Machine (LwM2M) protocols, enabling device-specific configuration and chaining of computational operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full firmware updates are used to update computational models on constrained devices, then the computational models can be updated, but large amounts of data must be transferred and the device must be rebooted
Solution Approach 1:
The patent segments the firmware update process into two distinct parts: (1) transferring only the computational model data object (containing the model itself, input data, and output data) and (2) configuring the device to execute the new model. This segmentation allows updating computational models without transferring entire firmware images, significantly reducing data transfer volume while maintaining update capability.
Solution Approach 2:
The patent extracts the computational model as a separate, independently transferable data object from the firmware. The model is configured as a distinct entity that can be transferred and executed without requiring full firmware replacement. This extraction enables selective updates of computational models while preserving the rest of the firmware, reducing data transfer requirements.
2Reliability
If full firmware updates are used to update computational models, then the computational models can be updated, but the device requires reboot which causes operational issues
Solution Approach 1:
The patent segments the update process so that only the computational model component is replaced, not the entire firmware. This allows the device to load and execute new models without requiring a full system reboot, maintaining operational continuity while still enabling model updates.
Solution Approach 2:
The patent configures the device to execute the new computational model in advance of any required system operations. By pre-configuring the model execution environment and loading mechanisms, the system can swap models without interrupting overall device operation, eliminating the need for disruptive reboots.
3Ease of manufacture
If generalized firmware updates are used across different device types, then updates can be applied broadly, but device-specific configuration and capabilities cannot be properly addressed
Solution Approach 1:
The patent creates a universal computational model data object structure that can be transferred to multiple device types through a common interface. The model definition, input data specifications, and output data specifications are formulated in a device-agnostic manner, allowing the same update mechanism to serve multiple device types while maintaining device-specific execution capabilities.
Solution Approach 2:
The patent enables device-specific configuration by allowing each device to interpret and execute the universal model data object according to its local capabilities and requirements. The computational model can be adapted to specific device contexts through local configuration parameters and execution environments, ensuring proper functionality across different device types without requiring separate update processes.
Data Source
AI summary
A computing node is disclosed. The computing node comprises processing circuitry configured to cause the computing node to receive a message (102) comprising configuration information for a resource of a data object that is hosted at the computing node and is associated with a computational operation, which computational operation is executable by the computing node. The processing circuitry is further configured to cause the computing node to configure (104) the resource of the data object on the computing node in accordance with the received configuration information, and to execute (106) the computational operation in accordance with the configured resource. Also disclosed are a corresponding server node and methods of operating a computing node and a server node. The computing node may comprise a Lightweight Machine to Machine (LwM2M) client and the server node may comprise an LwM2M server.