Time-Stamped Network Graphs for Topology, State, and Configuration Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network management systems struggle to efficiently compare network topology, states, and configuration at different time instances, particularly in complex and distributed networks, leading to cumbersome and error-prone troubleshooting processes.
Innovation Solution
A network management system represents the network as a Network Image (NI) graph, where entities are vertices and relationships are edges, with timestamp pairs indicating active periods, allowing efficient traversal to determine network topology, states, and configuration at specific time instances, and uses auxiliary graphs to capture changes efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional network management methods are used to compare network topology at different time instances, then troubleshooting can be performed, but the process becomes cumbersome and error-prone in complex distributed networks
Solution Approach 1:
The patent creates simplified copies of the complex network topology at different time instances using graph representations. These graph copies capture essential connectivity information without the full complexity of the original network, enabling reliable comparison while reducing management burden. The graph models serve as manageable representations that can be stored, retrieved, and compared efficiently.
Solution Approach 2:
The patent segments the network comparison process into distinct steps: capturing topology at specific time instances, representing them as graphs, storing with timestamps, and comparing only relevant portions. This segmentation transforms a cumbersome monolithic process into manageable discrete operations, reducing errors and improving reliability in troubleshooting.
2Productivity
If network topology changes are tracked in complex distributed networks, then troubleshooting efficiency improves, but the time and resources required for comparison increase
Solution Approach 1:
The patent performs preliminary actions by capturing and storing network topology as graph representations at specific time instances before troubleshooting is needed. These pre-captured graphs with timestamps are stored for later retrieval and comparison, eliminating the need for time-consuming live analysis during actual troubleshooting events.
Solution Approach 2:
By creating simplified graph copies of the network topology that capture essential connectivity information, the patent enables rapid comparison operations. These copies are much lighter than full network state representations, allowing efficient retrieval and comparison even in large distributed networks, thus improving troubleshooting efficiency without excessive time loss.
3Measurement precision
If detailed network configuration data is maintained for comparison, then accurate change identification is achieved, but storage requirements and processing overhead increase
Solution Approach 1:
The patent extracts only the essential connectivity information needed for topology comparison, representing it as graphs with vertices and edges. This extraction removes unnecessary detailed configuration data while retaining the critical topology structure, achieving accurate change detection with reduced storage requirements.
Solution Approach 2:
The patent applies local quality by representing different aspects of network topology with appropriate graph elements (vertices for network entities, edges for connections). This localized representation focuses storage and processing on relevant topology characteristics rather than maintaining all possible configuration details uniformly across the entire network.
Data Source
AI summary
A network management system for orchestrating a network is provided. During operation, the system generates a graph representing the network. A respective vertex corresponds to an entity in the network, and a respective edge indicates a relationship between a vertex pair. The system can determine a first and a second timestamps for a respective edge. The first timestamp indicates a time instance when a relationship indicated by the edge is established. The second timestamp indicates a time instance when the relationship is terminated. The time range between the first and second timestamps indicates an active period for the edge. The system then receives, from an interface of the system, an instruction for comparing the topology, states, and configuration of the network. The system determines the topology, states, and configurations of the network at a target time instance indicated by the instruction by traversing the active edges of the graph.