Service Dependency Mapping via Transfer Entropy Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current tools for identifying service dependencies in data networks are limited in tracing chains of dependencies beyond two services and fail to utilize all available data, leading to reduced accuracy in dependency identification.
Innovation Solution
A method and apparatus that utilize a connection manager, time series builder, and dependency evaluator to identify connections, create time series from monitoring data, and compute transfer entropy to discover service dependencies across multiple services in a data network, enabling passive detection without disrupting data flow.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current tools are used to identify service dependencies, then basic dependency identification is possible, but the ability to identify chains of dependencies across more than two services is lost
Solution Approach 1:
The patent segments the dependency identification process into distinct functional components: a connection manager that identifies connections between nodes, a time series builder that creates time series data from monitoring data, and a dependency evaluator that computes transfer entropy to determine dependencies. This segmentation enables the system to handle complex multi-service dependency chains by breaking down the analysis into manageable steps, thereby improving measurement precision without being overwhelmed by complexity.
Solution Approach 2:
The patent introduces a temporal dimension by creating time series from monitoring data and using transfer entropy computation that considers time-lagged relationships. This dimensional approach allows the system to trace dependency chains across multiple services by analyzing how changes in one service propagate through time to affect other services, enabling accurate identification of multi-service dependency chains that traditional static tools cannot detect.
2Measurement precision
If current tools are used for dependency identification, then some dependencies can be identified, but not all available data is utilized which reduces accuracy
Solution Approach 1:
The patent creates a universal framework that can process multiple types of monitoring data from various sensors simultaneously. The time series builder component aggregates data from numerous sources, and the dependency evaluator uses transfer entropy computation that can handle diverse data types. This multi-functional approach ensures that all available data is utilized effectively, improving measurement precision by comprehensively analyzing the entire dataset rather than selecting only specific portions.
Solution Approach 2:
The system implements feedback mechanisms where the dependency evaluator continuously computes transfer entropy based on time series data and refines dependency identification. The connection manager receives feedback about identified dependencies and adjusts connection identification accordingly. This feedback loop ensures that all available monitoring data is iteratively utilized to improve the accuracy of dependency identification, maximizing data utilization while enhancing measurement precision.
Data Source
AI summary
A method and apparatus for discovering service dependencies. A plurality of connections is identified between nodes in a data network. A set of connection pairs is identified based on the plurality of connections identified. A set of time series is created for the set of connection pairs using monitoring data received from a plurality of sensors monitoring the data network. Service dependencies may be discovered using the set of time series.