Distributed Protocol Analysis Using Two-Stage ML Variable Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for implementing system-wide protocols in distributed computing systems require extensive computational resources to aggregate and process data, which is inefficient for users who only need to evaluate a few metrics across a large number of network devices.
Innovation Solution
A system utilizing two machine learning engines, where a first engine identifies salient variables and provides them to secondary engines for data visualization on user devices, reducing computational and network resources while allowing user-friendly XR data visualization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods aggregate and process data across all network devices, then comprehensive system-wide analysis is achieved, but computational resources and network traffic increase significantly
Solution Approach 1:
The patent segments the monolithic data processing task into two distinct machine learning engines: a first ML engine that performs comprehensive system-wide analysis to identify salient variables, and second ML engines deployed at individual network devices that analyze only the specific variables of interest. This segmentation allows the system to maintain comprehensive analysis capabilities while reducing the computational burden at each device by processing only relevant data locally.
2Reliability
If conventional methods process all data across the distributed system, then complete protocol evaluation is achieved, but network traffic increases
Solution Approach 1:
The patent extracts only the essential salient variables identified by the first ML engine and transmits them to the second ML engines at individual devices, rather than transmitting or processing all raw data. This extraction approach maintains the reliability of protocol evaluation by ensuring that the most important variables are analyzed locally, while dramatically reducing network traffic by eliminating the need to transmit voluminous raw datasets.
3Loss of information
If a single centralized system processes all protocol data, then system-wide insights are obtained, but processing time increases
Solution Approach 1:
The first ML engine performs preliminary analysis of the protocol dataset across the entire system to identify which variables are most strongly correlated with protocol performance. This preliminary action creates a filtered set of salient variables that can then be quickly analyzed by multiple second ML engines in parallel at individual devices, significantly reducing the time required for local analysis while preserving system-wide insights through the initial comprehensive processing.
Data Source
AI summary
Systems, computer program products, and methods are described herein for intelligent protocol implementation in a distributed computing system. The present disclosure is configured to monitor usage data for a plurality of network devices; receive, from a first network device, a protocol dataset; analyze the protocol dataset using a first machine learning engine; determine, based on an output of the first machine learning engine, at least one variable correlated with the protocol dataset; and instruct a second machine learning engine to analyze usage data associated with the at least one variable, wherein the second machine learning engine is configured to generate a data visualization on a display of the first network device.