Common Metric Service for SDN Data Synchronization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Software-Defined Networks (SDNs) face challenges in processing metrics globally, as the controller is unaware of metrics tracked by external components, leading to incomplete data processing and synchronization issues across the network.
Innovation Solution
A common metric service is introduced that allows components both internal and external to the controller to register and process metrics using a metric proxy, enabling centralized processing and synchronization of metrics across the SDN, regardless of their location, and persisting summarized values as time series data for analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If the controller processes metrics independently without a common metric service, then the controller maintains simple architecture and fast local processing, but the controller cannot access metrics from external components leading to incomplete data processing
Solution Approach 1:
A common metric service is introduced as an intermediary component between external components and the controller. This metric service receives, processes, and standardizes metrics from external components, then makes them available to the controller. The controller interacts with the metric service through defined interfaces rather than directly accessing external components, ensuring complete metrics data while maintaining a manageable architecture through clear separation of concerns.
2Reliability
If external components track metrics independently, then each component maintains autonomy and simple local tracking, but synchronization issues arise across the network
Solution Approach 1:
The patent merges the metrics tracking and processing functions into a centralized common metric service that all components interact with. Instead of each component independently tracking metrics which causes synchronization issues, the metric service provides a unified processing mechanism that ensures all metrics are synchronized across the network while maintaining a standardized interface for ease of operation.
Solution Approach 2:
The metric service implements feedback mechanisms where metrics are continuously collected, processed, and made available to all relevant components. This feedback loop ensures that all components have access to the latest metrics data, maintaining synchronization across the network while the metric service handles the complexity of data management centrally.
3Productivity
If the controller is unaware of external component metrics, then the controller maintains clear boundaries and simple management, but global metric processing cannot be achieved
Solution Approach 1:
The common metric service is designed as a universal component that handles metrics from both internal and external components through a standardized interface. This multi-functional service can process different types of metrics from various sources, enabling global metric processing efficiency while the controller interacts with this universal service rather than directly integrating with each individual component, thus managing integration complexity.
Data Source
AI summary
Processing a metric of a component of a SDN can include registering a unique identifier associated with the metric in storage of the SDN.