Stitching Partial Network Topologies for Scalable Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Determining the topology of very large networks becomes impractical due to exponential time requirements, making real-time or near-time network analysis and problem detection infeasible, especially when using existing topology inference tools that are computationally intensive and not scalable for networks with tens of thousands of devices.
Innovation Solution
The method involves determining individual network topologies for smaller subsets of a network, filtering out internal connection information to create residual information, and then using this residual information to stitch together the topologies of multiple networks, thereby avoiding the need to process all devices and connections within the network, allowing for parallel processing and reducing processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If topology inference tools process all devices in a very large network, then comprehensive network topology determination is achieved, but processing time increases exponentially making real-time analysis infeasible
Solution Approach 1:
The patent divides a very large network into multiple smaller subnetworks, each managed by separate network management systems. Each NMS determines topology for its own subnetwork independently, avoiding the exponential processing time required to analyze the entire network as a single unit. The topology data from multiple NMSs is then merged to provide a comprehensive view of the entire network.
2Productivity
If multiple network management systems manage different subnetworks, then processing time is reduced through parallel processing, but a consolidated network topology view is not provided to administrators
Solution Approach 1:
The patent merges topology data from multiple network management systems into a consolidated network topology view. Each NMS provides its determined topology for its subnetwork, and these individual topologies are combined to create a comprehensive network topology that provides administrators with a complete view of the entire network while maintaining the processing efficiency of parallel operations.
3Area of stationary object
If network size increases, then network capacity and coverage are improved, but the complexity and time required for topology determination increases exponentially
Solution Approach 1:
The patent segments the large network into smaller subnetworks that can be managed independently. This segmentation reduces the complexity of topology determination for each individual NMS, as each system only needs to analyze a subset of devices rather than the entire network. The overall network coverage remains large, but the computational complexity is distributed across multiple smaller analysis tasks.
Data Source
AI summary
A method and system that takes advantage of processes that are efficient for determining the topology of small to medium size networks to determine individual network topologies for such networks, and then merges these individual topologies into a consolidated topology for the entire network. Each of the processes that determines the topology of the smaller networks provides the determined network topology, as well as a list of factors that may be relevant in the determination of how the given topology might be attached to any other given topology, such as the identification of a node that is not included in the given topology, or other indications of external connections. The merging process is configured to substantially restrict its analysis to these factors, thereby limiting the extent, and therefore the time consumed, by this stitching and merging process.