Offline Organizational Network Analysis Using Distributed Graph Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing organizational network analysis (ONA) tools face performance issues with large user networks due to memory and processor intensity, and privacy concerns arise from data sharing, limiting their ability to handle datasets of thousands of users effectively and efficiently.
Innovation Solution
A system utilizing distributed computing platforms for offline analysis using parallelized graph metrics, enabling larger network analysis and addressing privacy concerns by processing data within the organization's computing system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If online real-time ONA tools are used to analyze organizational networks, then the analysis can be performed interactively, but the system becomes unusable or impossible with more than a few thousand users due to memory and processor intensity
Solution Approach 1:
The patent divides the organizational network analysis into discrete graph metric calculations that can be processed in parallel. The system segments the large-scale network analysis into individual graph metric computations for multiple users, which are then executed concurrently using distributed computing resources, enabling both interactivity and scalability to thousands of users.
Solution Approach 2:
The patent introduces an intermediary processing layer that receives network data, performs parallel graph metric calculations using distributed computing platforms, and returns results. This intermediary layer acts as a buffer between the user interface and the computationally intensive analysis, enabling interactive operation while handling large networks through offline parallel processing.
2Measurement precision
If data is shared with external ONA tools for analysis, then comprehensive network metrics can be calculated, but privacy restrictions and regulatory concerns prevent data sharing
Solution Approach 1:
The patent extracts the graph metric calculation algorithms from external ONA tools and implements them directly within the organization's computing environment. By taking out the computational logic and embedding it locally, the system achieves comprehensive network metrics analysis while keeping all organizational data within the organization's infrastructure, eliminating privacy and regulatory concerns.
Solution Approach 2:
The patent enables the organization's computing system to perform graph metric calculations independently without requiring external ONA tools. The system uses distributed computing platforms running locally to calculate network metrics, allowing the organization to serve its own analysis needs while maintaining data privacy and regulatory compliance.
3Quantity of substance
If graph metrics are calculated in serial for large networks, then memory and processor requirements are manageable, but computational wait times become unreasonably long
Solution Approach 1:
The patent segments the calculation of graph metrics into independent parallel tasks that can be executed simultaneously. Instead of calculating metrics for each user in serial order, the system divides the computation into discrete units that run concurrently across multiple processing cores or distributed computing nodes, dramatically reducing total computation time while maintaining manageable resource requirements through parallel execution.
Solution Approach 2:
The patent implements dynamic parallel processing that adapts to the size and structure of the organizational network. The system dynamically distributes calculation tasks across available computing resources, adjusting the degree of parallelization based on network characteristics and available hardware, thereby optimizing both computation time and resource utilization.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method may include presenting a graphical user interface (GUI) configured to define a network for analysis, the GUI including: a first portion configured to define a network boundary of an organizational network graph (ONG) in a dataset; a second portion configured to define a beginning time and ending time for the network analysis; and a third portion configured to define a set of one or more graph metrics for vertices in the ONG; retrieving a portion of the dataset based on the network boundary, beginning time, and ending time; transmitting instructions to a distributed computing platform to generate the set of graph metrics on the portion of the dataset; and storing the generated set of graph metrics as associated with the network analysis.