Predictive Network Topology Subsets for Cloud Failure Remediation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network providers face challenges in comprehending the relationships between entities in complex cloud networks, leading to difficulties in optimal visualization and management, even with orchestration systems like Kubernetes.
Innovation Solution
A system that predicts future anomalous characteristics and failures in network topologies based on historical data, generating recommendations to prevent such issues by modifying network topology portions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If orchestration systems like Kubernetes are used to deploy and manage cloud networks, then network deployment and management become more automated, but network providers still cannot comprehend the relationships between entities of the network
Solution Approach 1:
The patent introduces an intermediary system that sits between the orchestration system and network providers. This intermediary automatically discovers, tracks, and visualizes relationships between network entities (services, nodes, pods, containers) that orchestration systems manage but do not inherently expose. The intermediary captures metadata, builds relationship graphs, and presents comprehensible visualizations, thereby bridging the information gap without reducing automation.
2Loss of information
If network providers attempt to visualize and manage all entities in complex cloud networks, then complete network comprehension is achieved, but the complexity of visualization and management increases significantly
Solution Approach 1:
The patent segments the complex network visualization into manageable subsets based on relationships to selected entities. Instead of displaying all network entities simultaneously, the system allows users to select anchor entities and automatically displays only the relevant subset of entities and relationships connected to those anchors. This segmentation reduces visual complexity while preserving complete relationship information for the selected scope.
Solution Approach 2:
The patent applies local quality by providing different levels of detail and interaction for different parts of the network visualization. Selected anchor entities receive prominent display with detailed relationship information, while other entities are displayed with appropriate detail based on their relevance. The system allows deep inspection of specific relationships while maintaining an overview of the broader network context.
3Loss of information
If network providers manually track relationships between network entities, then complete comprehension is possible, but the time and resources required increase significantly
Solution Approach 1:
The patent implements self-service by enabling the system to automatically discover, track, and update relationships between network entities without manual intervention. The system continuously monitors the cloud network, automatically builds and updates relationship graphs, and maintains current visualizations. This automated self-service approach provides complete relationship comprehension while eliminating the time and resources that would be required for manual tracking.
Data Source
AI summary
Techniques for recommending plans to remediate a network topologies are disclosed. The techniques include predicting characteristics of the network using network topology information identifying relationships between entities in the network. The techniques further include determining a subset of the topology based on the predicted characteristics violating anomaly detection criteria. Additionally, the techniques include determining a remediation plan for modifying the subset and presenting the plan to a user.