Real-Time Dataflow Tools for Low-Latency Edge Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing enterprise software systems face challenges in handling large volumes of data generated by industrial machines due to connectivity issues, high latency, and cost-prohibitive bandwidth, which hinder real-time decision-making and predictive maintenance.
Innovation Solution
A dataflow programming language with accompanying tools for edge computing, enabling real-time data analysis and pattern-driven intelligence at the source of IoT data, using a graphical user interface for program development, and a software development kit for edge app creation, along with a cloud-based management console for deployment and analytics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If data is sent to cloud storage for processing, then centralized computing power and storage are utilized, but bandwidth requirements increase, latency increases, and connectivity dependencies arise
Solution Approach 1:
The patent segments the centralized cloud computing architecture into distributed edge computing nodes deployed at multiple locations closer to data sources. This segmentation allows computing power to be distributed rather than concentrated in remote data centers, reducing the bandwidth and energy required to transmit data across long distances while maintaining access to computational resources.
Solution Approach 2:
The patent introduces edge computing devices as intermediary systems between data sources and cloud storage. These intermediaries perform local data processing and filtering, reducing the volume of data that needs to be transmitted to the cloud and thereby reducing bandwidth consumption and energy loss while still leveraging centralized cloud resources when needed.
2Power
If data is sent to cloud storage for processing, then centralized computing resources are utilized, but latency increases due to distance and network variability
Solution Approach 1:
The patent segments the centralized cloud architecture into distributed edge computing nodes positioned geographically closer to data sources and end users. This segmentation reduces the physical distance data must travel, thereby reducing network latency and variability while still providing access to computing resources through the distributed network of edge nodes.
Solution Approach 2:
The patent implements preliminary data processing and filtering at edge computing nodes before data is transmitted to cloud storage. This preliminary action reduces the volume of data requiring long-distance transmission and enables faster initial processing, thereby reducing overall latency while still utilizing centralized cloud resources for more complex computations.
3Loss of information
If all sensor data is transmitted to cloud storage, then complete data availability is achieved, but bandwidth costs become prohibitive
Solution Approach 1:
The patent extracts and processes data locally at edge computing nodes, filtering and preprocessing data before transmission to cloud storage. This extraction of essential information at the source reduces the volume of data requiring expensive bandwidth for transmission while maintaining availability of critical information through local processing and selective cloud storage.
Solution Approach 2:
The patent implements partial data transmission to the cloud, sending only processed, filtered, or aggregated data rather than all raw sensor data. This partial action approach maintains sufficient data availability for cloud-based analysis while dramatically reducing bandwidth consumption and associated costs.
4Loss of time
If edge computing is implemented, then real-time data processing and reduced latency are achieved, but device complexity increases
Solution Approach 1:
The patent implements standardized edge computing devices with multi-functional capabilities that can perform various data processing tasks locally. These universal devices reduce the need for specialized hardware at each location, managing device complexity through standardization while still enabling real-time processing through distributed intelligence.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A dataflow programming language can be used to express reactive dataflow programs that can be used in pattern-driven real-time data analysis. One or more tools are provided for the dataflow programming language for checking syntactic and semantic correctness, checking logical correctness, debugging, translation of source code into a secure, portable format (e.g., packaged code), translation of source code (or packaged code) into platform-specific code, batch-mode interpretation, interactive interpretation, simulation and visualization of the dataflow environment, remote execution, monitoring, or any combination of these. These tools embody a method of developing, debugging, and deploying a dataflow graph device.