Datacenter Network Topology Identification Without Configuration Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current datacenter network monitoring tools lack the ability to visualize the physical network topology, such as leaf-spine architecture, without relying on configuration data, which is often unavailable.
Innovation Solution
A method that uses connection information between network elements to identify and visualize groups of network elements organized in specific architectures like leaf-spine, by assigning layers based on distance from host computers and using a connection matrix to determine direct connections between layers, allowing for the identification and amalgamation of groups.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If network monitoring tools use configuration data to identify physical network topology, then network architecture identification accuracy is improved, but system complexity increases and configuration data availability becomes unreliable
Solution Approach 1:
The network elements themselves generate and provide the connection information needed for topology identification. Each network element autonomously reports its connections to monitoring tools, eliminating the need for external configuration data collection and processing systems.
Solution Approach 2:
The patent replaces the mechanical/manual process of collecting and parsing configuration files with an automated information gathering system where network elements directly report their connection status through standardized protocols, eliminating complex file processing mechanisms.
2Loss of information
If network monitoring tools extract topology from configuration data, then network layout visualization is achieved, but reliability decreases due to configuration data unavailability
Solution Approach 1:
Network elements continuously report their connection status to monitoring tools, creating a real-time feedback mechanism that ensures topology information is always available and current, eliminating reliance on static configuration files that may be missing or outdated.
Solution Approach 2:
The system proactively collects connection information from network elements before topology identification is needed, ensuring that the necessary data is already available and validated, rather than attempting to extract it from configuration files only when required.
3Adaptability or versatility
If connection information is gathered from traffic monitoring, then configuration data dependency is reduced, but data collection complexity increases
Solution Approach 1:
The patent introduces standardized information reporting protocols as intermediaries between network elements and monitoring tools. These protocols simplify the data collection process by providing a uniform interface, eliminating the need for complex traffic analysis while maintaining independence from configuration files.
Data Source
AI summary
Some embodiments provide a method for identifying network architecture in a datacenter. The method identifies connection information for multiple network elements. The connection information for each network element specifies other network elements to which the network element connects. The method assigns a set of the network elements to a set of two or more layers based on the connection information for the network element. The method uses the connection information to identify at least one group of network elements organized into a particular network architecture. The identified groups are for use in displaying user interface visualizations of the network elements of the datacenter.