Network Observation Consistency via Cloud Mediator and Noise Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In computer network environments, making reliable observations is challenging due to factors like network instability, device misbehavior, communication failures, and noise, which complicates data recording and noise elimination.
Innovation Solution
A system comprising a cloud server and agents that periodically scan the network, using machine learning to correct observations by statistically analyzing data, determining confidence levels, and maintaining an environment model to filter out noise and provide consistent device properties.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If agents scan the network to collect observations, then data coverage is improved, but noise and faulty observations increase
Solution Approach 1:
A cloud server acts as an intermediary between multiple scanning agents and the final observation results. The cloud server receives observations from multiple agents, applies statistical analysis to identify consistent patterns, and filters out noisy or faulty observations. This mediator approach allows comprehensive data collection while systematically eliminating noise through cross-validation of multiple independent observations.
Solution Approach 2:
The system implements feedback mechanisms where observation results are continuously analyzed and used to improve future scanning. The cloud server tracks observation consistency over time, identifies sources of noise, and adjusts scanning parameters or agent behavior based on accumulated data. This feedback loop enables the system to learn from past noisy observations and improve filtering effectiveness.
2Reliability
If multiple agents are deployed to improve observation reliability, then data consistency should improve, but system complexity increases
Solution Approach 1:
Multiple independent agent instances are merged into a unified system managed by a central cloud server. While individual agents remain simple and independent for ease of deployment, their combined observations are processed collectively to achieve reliability through statistical analysis. The merging occurs at the data aggregation and analysis level, not at the agent level, thus maintaining agent simplicity while achieving system-level reliability.
Solution Approach 2:
The cloud server performs multiple functions: receiving observations from agents, storing historical data, performing statistical analysis, filtering noise, and generating corrected observations. This multi-functional approach consolidates complexity into a single component rather than distributing it across multiple agents, simplifying the overall system architecture while maintaining high reliability through comprehensive data processing.
3Measurement precision
If statistical analysis is applied to correct observations, then measurement precision is improved, but processing time increases
Solution Approach 1:
The cloud server performs preliminary statistical analysis on accumulated observations to establish baseline patterns of device behavior and typical observation values. This preliminary analysis creates reference data that can be quickly compared against new observations, reducing the computational burden of real-time analysis. By pre-processing historical data, the system prepares lookup tables and statistical models that accelerate subsequent correction operations.
Solution Approach 2:
The system applies statistical analysis selectively rather than uniformly to all observations. It focuses computational resources on observations that show inconsistency or potential faults, while accepting observations that clearly match established patterns without extensive analysis. This partial application of rigorous statistical methods reduces overall processing time while maintaining precision for critical cases.
Data Source
AI summary
A system persistently presents consistent properties of devices on a local network based on observed values for the network devices, and values derived from the observed values. The observed values may be received from an agent based on scans of the local network. Even though some scans may be faulty resulting in missing or incorrect data, a user can be consistently presented with properties of the device, even when the missing or incorrect data would otherwise cause a change to the property. For instance, the system may replace a data value that is either missing or determined to be incorrect with a value that is determined, based on historical or lab observations, to be the likely correct value based on the assumed state or likely state of the observed device.