Action Recommendation Engine for Network Operations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current software products in Network Operations Centers (NOCs) face challenges in providing effective guidance for network actions due to reliance on expert rules, which become difficult and expensive for complex scenarios, especially in multi-vendor or multi-domain environments, and require time-consuming and resource-intensive processes to determine network states.
Innovation Solution
An Action Recommendation Engine (ARE) that receives raw, unprocessed data from network elements to directly associate with remedial actions, bypassing the need to explicitly determine network states, and utilizes Machine Learning (ML) and Reinforcement Learning (RL) to recommend actions, enabling faster and more effective network management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If expert rules are used to provide guidance for network actions, then some partial compensation can be provided, but the process becomes time-consuming and resource-intensive
Solution Approach 1:
The system enables self-service by allowing the network operations center to automatically determine network states and generate action recommendations without requiring extensive external expert intervention. The machine learning model learns from historical data to autonomously identify patterns and suggest actions, reducing dependency on manual expert analysis while maintaining reliable guidance quality.
2Reliability
If expert rules are used for complex scenarios, then guidance can be provided, but determining effective rules becomes incrementally difficult and expensive
Solution Approach 1:
The patent replaces the mechanical system of manual expert rule creation with an automated machine learning system. Instead of experts manually formulating complex rules for each scenario, the system automatically learns patterns from historical network data and generates action recommendations, significantly reducing the complexity and cost associated with handling complex scenarios.
Solution Approach 2:
The system changes the approach from static expert rules to dynamic machine learning models that adapt to varying network conditions. By training on historical data with multiple parameters and features, the system can handle complex scenarios by automatically adjusting its decision-making based on learned patterns rather than relying on pre-defined complex rules.
3Adaptability or versatility
If expert rules are used for multi-vendor or multi-domain scenarios, then collective domain expertise can be codified, but the process becomes incrementally difficult and expensive
Solution Approach 1:
The machine learning system provides universality by learning from diverse historical data across multiple vendors and domains simultaneously. Rather than creating separate rule sets for each vendor or domain, the system identifies common patterns and anomalies across all data sources, enabling it to handle multi-vendor and multi-domain scenarios with a single unified model that adapts to different contexts.
4Measurement precision
If network state determination is performed explicitly, then input for ARE can be provided, but the process becomes difficult or expensive
Solution Approach 1:
The patent extracts and eliminates the need for explicit network state determination as a separate step. Instead of requiring the system to explicitly identify and classify network states before generating recommendations, the machine learning model directly processes raw network data and outputs action recommendations, removing the complex intermediate step of state determination while maintaining the necessary accuracy through learned patterns.
Data Source
AI summary
Systems, methods, and computer-readable media are provided for recommending actions to be taken in a network for optimizing or improving the operability of the network. A method, according to one implementation, includes a first step of receiving raw, unprocessed data that is obtained directly from one or more network elements of a network. The method includes second step of determining one or more remedial actions using a direct association between the raw, unprocessed data and the one or more remedial actions.