Tiered IoT Data Processing with Local and Server Model Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

IoT devices face challenges in maintaining prediction accuracy due to limited computational power and changing data conditions, leading to reduced reliability of local predictions compared to more powerful server-based models.

Innovation Solution

Implementing a split prediction system where IoT devices (hub or edge devices) process data locally and transmit it to a provider network for correction using more accurate models, allowing for periodic updates of local models based on new data to enhance prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a simple data processing model is implemented on an IoT device, then the device can perform local data analysis and generate predictions, but the reliability and accuracy of predictions deteriorate compared to larger models on more powerful computing devices

Engineering Contradiction:
Improvelocal data processing capabilityVSAvoidprediction accuracy
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system segments the data processing function into two parts: a simple local model on the IoT device for immediate data analysis, and a larger remote model on the provider network for enhanced accuracy. The local model handles real-time processing while the remote model provides periodic corrections and updates, resolving the contradiction between local capability and prediction reliability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system implements feedback by transmitting local prediction results and new data to the provider network, which then generates model updates based on the performance gap between local and remote models. These updates are transmitted back to the IoT device to improve the local model's accuracy over time, maintaining reliability while preserving local processing benefits.

Inventive Principle:
Principle #23Feedback

2Reliability

If a larger data processing model with more parameters is used, then prediction accuracy improves, but the computational power and processing speed required increase

Engineering Contradiction:
Improveprediction accuracyVSAvoidcomputational power requirement
Core Design Contradiction:
ReliabilityVSPower

Solution Approach 1:

The computational workload is segmented between the IoT device and the provider network. The IoT device runs a lightweight local model that requires minimal computational power for real-time processing, while the provider network hosts the computationally intensive larger model that performs periodic training and generates model updates, thus distributing the power requirements appropriately.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of deploying the full large model to the IoT device, the system creates a simplified copy or approximation that runs locally. This local model is then periodically updated with corrections and adjustments generated by the larger remote model, achieving high accuracy without requiring the full computational resources of the original large model on the constrained device.

Inventive Principle:
Principle #26Copying

3Reliability

If the local model on an IoT device is updated periodically with new data, then prediction accuracy improves, but the complexity of model management and data transmission increases

Engineering Contradiction:
Improvemodel accuracy over timeVSAvoidmodel update management
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The provider network acts as an intermediary that manages the complexity of model updates. It collects new data from the IoT device, trains or fine-tunes models, generates updated model parameters, and transmits them back to the device. This intermediary approach simplifies the IoT device's role to just receiving and applying updates, rather than managing the entire model training and update process.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements self-service by having the provider network automatically perform model training, evaluation, and update generation based on incoming data from the IoT device. The update mechanism is designed to be automated, with the network service detecting when updates are needed, performing the training, and pushing updates without requiring manual intervention, thus managing complexity centrally while keeping the device simple.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP3893471A1Tiered data processing for IoT device clusters
Publication Date: 2021.10.13 AMAZON TECH INC
  • EP3893471A1 patent drawingFigure 1
  • EP3893471A1 patent drawingFigure 2
  • EP3893471A1 patent drawingFigure 3

AI summary

Edge devices of a network collect data. An edge device may determine whether to process the data using a local data processing model or to send the data to a tier device. The tier device may receive the data from the edge device and determine whether to process the data using a higher tier data processing model of the tier device. If the tier device determines to process the data, then the tier device processes the data using the higher tier data processing model, generates a result based on the processing, and sends the result to an endpoint (e.g., back to the edge device, to another tier device, or to a control device). If the tier device determines not to process the data, then the tier device may send the data on to another tier device for processing by another higher tier model.