Unified Graph Models for Cross-Layer Network Entity Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network analysis systems fail to provide a unified, comprehensive view of both explicit and implicit network components across multiple layers, leading to inefficient analysis and increased costs due to separate representation of similar objects and lack of integration with cloud platform host systems.
Innovation Solution
A method and system for generating unified graph models that collect, genericize, and store network entity data features to create a multi-dimensional data structure representing network entities and their relationships, including both explicit and implicit objects across various layers, using a graph database for storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traffic-independent solutions are used to identify environment components and connections, then comprehensive understanding of network components is improved, but the ability to provide simple unified views is worsened
Solution Approach 1:
The patent merges multiple similar network components (firewalls, load balancers, routers) from different cloud providers into a single unified graph model. This combines scattered component representations into one comprehensive view, resolving the contradiction by maintaining completeness while simplifying the overall structure through unified representation.
Solution Approach 2:
The unified graph model serves multiple functions simultaneously: it represents components from different cloud providers, models connections across layers, enables graph-specific queries, and provides simplified visualization. This multi-functional approach allows the system to maintain comprehensive understanding while delivering simple unified views through a single versatile data structure.
2Measurement precision
If separate analysis and representation of each object is performed, then detailed analysis capability is improved, but analysis efficiency is worsened
Solution Approach 1:
The patent combines separate analysis of individual network components into a unified graph model that represents all components and their relationships simultaneously. This merging enables graph-specific commands and queries to operate across the entire network environment in one operation, dramatically improving analysis efficiency while maintaining the ability to examine individual components in detail when needed.
Solution Approach 2:
The patent transitions from analyzing components in isolation (one-dimensional) to representing them in a multi-dimensional graph structure that captures relationships, connections, and hierarchies across different layers. This dimensional change enables simultaneous analysis of multiple components and their interactions, improving efficiency without sacrificing detailed analysis capability.
3Measurement precision
If multiple similar network components are represented separately, then individual component accuracy is improved, but representation conciseness is worsened
Solution Approach 1:
The patent merges multiple similar network components into a unified graph model where each component is represented as a node with standardized properties. This merging maintains accurate representation of individual components through their node attributes while achieving conciseness through the unified graph structure that represents all components in a single compact model rather than separate scattered representations.
4Reliability
If traffic analysis systems are used to identify network activity, then network activity detection is improved, but representation of inactive components is worsened
Solution Approach 1:
The patent segments the network environment representation into two complementary parts: traffic analysis data that captures active network communications, and inventory data from cloud provider APIs that captures all components including inactive ones. By segmenting the data sources and combining them in a unified graph model, the system maintains reliable activity detection while ensuring complete representation of all network components regardless of their current activity state.
Data Source
AI summary
A system and method for generation of unified graph models for network entities are provided. The method includes collecting, for at least one network entity of a plurality of network entities, at least one network entity data feature, wherein the at least one network entity data feature is a network entity property; genericizing the collected at least one network entity; generating at least a network graph, wherein the generated network graph is a multi-dimensional data structure providing a representation of the plurality of network entities and relations between the network entities of the plurality of network entities; and storing the generated at least a network graph.