Edge Node Parallel Thread Architecture for Low Latency Analytics

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current computing architectures for edge devices are limited by their computing resources, leading to high latency in processing and analysis, particularly in applications like computer vision, which require rapid response times and are hindered by the limitations of processor capabilities and memory in edge computing devices compared to traditional computing systems.

Innovation Solution

A computing architecture that leverages edge nodes with multiple independent threads processing input data streams in parallel, utilizing both CPU and GPU for rapid data processing and analysis, enabling machine learning models to perform low-latency analytics and control functions by distributing processing tasks across multiple processors and caching data for quick access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If traditional centralized computing devices are used for data processing, then computing power is sufficient, but network latency increases and response time becomes unacceptable

Engineering Contradiction:
Improveresponse timeVSAvoidnetwork latency
Core Design Contradiction:
SpeedVSLoss of time

Solution Approach 1:

The system segments computing functions by deploying edge computing nodes at network perimeters closer to data sources. These nodes perform local data processing and filtering, while only transmitting essential results to centralized cloud infrastructure. This segmentation reduces network latency by eliminating unnecessary round-trips to distant centralized servers, directly resolving the contradiction between sufficient computing power and acceptable response time.

Inventive Principle:
Principle #1Segmentation

2Speed

If edge computing devices are used for local processing, then network latency is reduced, but computing resources are limited

Engineering Contradiction:
Improveprocessing speedVSAvoidcomputing resources
Core Design Contradiction:
SpeedVSPower

Solution Approach 1:

The patent introduces a hybrid architecture where edge computing nodes act as intermediaries between data sources and centralized cloud infrastructure. These nodes perform initial data filtering, preprocessing, and local analytics using their limited computing resources, then transmit only essential results to the cloud. This intermediary approach allows the system to leverage both the low-latency local processing capability of edge devices and the abundant computing resources of centralized cloud infrastructure, resolving the contradiction between processing speed and computing resource availability.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If multiple processing threads are implemented in parallel, then data processing throughput increases, but system complexity increases

Engineering Contradiction:
Improvedata processing throughputVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system segments processing tasks into distinct parallel threads executed by separate CPU and GPU units. Each thread handles specific processing functions (e.g., data reception, preprocessing, model evaluation, control signal generation), allowing independent optimization and management. This segmentation enables high-throughput parallel processing while maintaining manageable system complexity through modular task distribution across specialized processing units.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4152272B1Systems and methods for low latency analytics and control of devices via edge nodes and next generation networks
Publication Date: 2024.09.18 ACCENTURE GLOBAL SOLUTIONS LTD
  • EP4152272B1 patent drawingFigure 1
  • EP4152272B1 patent drawingFigure 2
  • EP4152272B1 patent drawingFigure 3

AI summary

A computing architecture providing for rapid analysis and control of an environment via edge computing nodes is disclosed. Input data streams may be captured via one or more data stream independent CPU threads and prepared for processing by one or more machine learning models. The machine learning models may be trained according to different use cases to facilitate a multi-faceted and comprehensive analysis of the input data. The evaluation of the input data against the machine learning models may be facilitated via independent GPU threads (e.g., one thread per model or use case) and the outputs of the models may be evaluated using control logic to produce a set of outcomes and control data. The control data may be utilized to generate one or more command messages that may provide feedback to a remote device or user regarding a state of a monitored environment or other observed condition.