Network Topology Inference via Hierarchical VLAN Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network topology determination methods are computationally demanding and inefficient, especially in large networks, as they require exhaustive comparisons of address forwarding tables and do not effectively handle incomplete or missing connectivity information, including the presence of virtual local area networks (VLANs) and aggregate links.
Innovation Solution
The method partitions the network into independent VLAN partitions and further partitions each VLAN based on nodes simply connected to selected root switches, using aggregate address forwarding tables and ancillary information like spanning tree or CDP data to reduce combinatorial complexity and validate inferences, thereby determining the network topology more efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exhaustive comparison of address forwarding tables is performed to determine network topology, then complete topology information can be obtained, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent segments the network into multiple partitions based on VLAN identifiers and other criteria, allowing topology determination to be performed independently on each partition rather than requiring exhaustive comparison across the entire network. This reduces computational complexity while maintaining accuracy within each partition.
Solution Approach 2:
The patent performs preliminary filtering of address forwarding table entries by VLAN identifier and other criteria before conducting topology comparisons. This preliminary action eliminates irrelevant data early in the process, reducing the computational burden of subsequent topology determination operations.
2Measurement precision
If exhaustive comparison of address forwarding tables is performed to determine network topology, then complete topology information can be obtained, but processing time increases significantly
Solution Approach 1:
By dividing the network into smaller partitions based on VLANs and other segmentation criteria, the patent enables parallel processing of topology determination across multiple partitions, significantly reducing overall processing time while maintaining complete topology coverage.
Solution Approach 2:
The patent applies partial action by performing topology determination only on relevant partitions and filtering out irrelevant address forwarding table entries, rather than exhaustively processing all possible combinations across the entire network.
3Measurement precision
If traditional topology determination methods are used, then complete topology can be determined, but the methods cannot effectively handle incomplete or missing connectivity information
Solution Approach 1:
The patent prepares for incomplete information by implementing validation mechanisms and alternative determination methods in advance. When connectivity information is missing or incomplete, the system can fall back on inference based on available data and validation against known network constraints, ensuring reliable topology determination even with incomplete input.
Solution Approach 2:
The patent implements feedback mechanisms where topology determination results are validated and used to refine subsequent processing. Incomplete or ambiguous information triggers additional validation steps and cross-checks, allowing the system to reliably determine topology even when initial data is incomplete.
4Device complexity
If network is partitioned into smaller parts, then computational complexity is reduced, but ensuring all nodes are included in topology becomes more difficult
Solution Approach 1:
The patent creates a universal topology determination framework that processes multiple partitions using the same methods and criteria. Each partition is processed independently but using consistent logic, and the results are aggregated to form the complete network topology, ensuring all nodes are included while maintaining computational efficiency.
Solution Approach 2:
The patent merges the topology results from individual partitions into a comprehensive network topology. By combining partition-specific topologies while validating for consistency and completeness, the system ensures all nodes are included in the final topology determination.
Data Source
AI summary
A network is partitioned into a set of independent partitions, and the topology of each partition is determined, then merged to form a topology of the entire network. Preferably, the partitioning is hierarchical, wherein the network is partitioned to form individual VLAN partitions, and each of the VLAN partitions is further partitioned based on the nodes that are simply connected to each port of one or more selected root switches within the VLAN partition. Simple connections to each port are efficiently determined based on an aggregate address forwarding table associated with each node. Ancillary information, such as spanning tree or CDP data, may be used to facilitate efficient partitioning and/or to validate inferences that are made with incomplete information.