Temporal Graph Network Analytics for Telemetry Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage solutions for network telemetry information are unable to simultaneously address snapshot retrieval, time traversal, and data retrieval across periods, leading to incomplete analysis and lack of modification reflection in stored data.
Innovation Solution
The implementation of temporal graphs that store time information with each vertex and edge, allowing for efficient traversal and retrieval of historical data, enabling snapshot retrieval and temporal analysis by maintaining multiple versions of graph elements with associated timestamps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If snapshot storage and retrieval system is used, then snapshot retrieval capability is improved, but time traversal functionality and data correlation across periods are lost
Solution Approach 1:
The patent segments the storage system into multiple independent components: snapshot storage for point-in-time retrieval, continuous data storage for time series analysis, and a temporal graph database for correlating events across time. Each component handles specific retrieval needs, allowing the system to provide both snapshot retrieval and time traversal capabilities simultaneously without compromise
Solution Approach 2:
The patent creates a multi-functional storage architecture where a single system provides diverse retrieval capabilities: point-in-time snapshots, continuous time-series data access, and temporal event correlation. The temporal graph database serves as a universal interface that can answer various types of temporal queries (snapshots, trends, correlations) through a unified data model
2Adaptability or versatility
If continuous data storage is implemented, then time traversal capability is improved, but snapshot retrieval efficiency decreases
Solution Approach 1:
The patent implements preliminary action by pre-computing and storing temporal indexes, event metadata, and graph relationships in advance. The temporal graph database pre-organizes continuous data into event streams with temporal annotations, allowing rapid snapshot retrieval without scanning entire continuous data sets during query execution
3Device complexity
If traditional storage solutions are used, then device complexity is reduced, but data correlation across spatial and temporal domains becomes impossible
Solution Approach 1:
The patent introduces a temporal graph database as an intermediary layer between traditional storage systems and analytical workloads. This intermediary transforms raw storage data into temporally annotated graph structures, enabling complex spatial-temporal correlations without requiring changes to underlying storage infrastructure, thus balancing complexity and capability
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for performing network analysis comprising: receiving telemetry information for a computer network, the telemetry information comprising network device state information for internetworking devices in the computer network and network traffic flow information for data packet flows through the computer network; transforming the telemetry information into a temporal graph that is digitally stored in computer memory; receiving, from a client computing device, a request for network analytics information for the computer network; in response to receiving the request for network analytics information, generating network analytics information for the computer network based on the temporal graph and sending the network analytics information to the client computing device.