Machine Learning Policy Assignment for Dynamic Application Loads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing load balancing mechanisms in networked computing environments struggle with dynamic and abrupt changes in workload distribution, leading to performance degradation and access difficulties due to inefficient assignment of policies across computing clusters.
Innovation Solution
Utilizing a machine learning model trained on multiple factors, including policy holder details, environmental events, and performance metrics to dynamically adjust the assignment of insurance policies across application instances, allowing for adaptive load balancing and resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional load balancing mechanisms are used to assign policies across computing clusters, then the system structure remains simple and easy to implement, but the system cannot adapt to dynamic and abrupt changes in workload, leading to performance degradation
Solution Approach 1:
An ML-based policy assignment service is introduced as an intermediary component between the computing clusters and the workload. This service receives workload requests, predicts the most suitable computing cluster using machine learning models, and assigns policies accordingly. The intermediary handles the complexity of dynamic adaptation internally while presenting a simplified interface to the rest of the system, thus improving adaptability without significantly increasing overall system complexity.
Solution Approach 2:
The patent replaces traditional rule-based or static load balancing mechanisms with machine learning-based dynamic prediction. Instead of using fixed algorithms or simple round-robin assignments, the system employs trained ML models that analyze historical workload data, environmental factors, and system state to intelligently predict optimal cluster assignments. This substitution enables the system to adapt to abrupt workload changes and environmental events while maintaining manageable complexity through automated learning rather than complex manual rule management.
2Productivity
If dynamic load balancing is implemented to handle abrupt workload changes, then system performance and adaptability improve, but the complexity of assignment mechanisms increases
Solution Approach 1:
The system performs preliminary actions by pre-training machine learning models using historical workload data, environmental event data, and performance metrics before actual workload assignment occurs. The ML models are trained offline to learn patterns and relationships in the data, so that when real-time workload assignment is needed, the models can quickly predict optimal assignments without requiring complex real-time calculations. This preliminary training phase separates the complexity of pattern recognition from the real-time assignment process, improving system performance while keeping the operational complexity manageable.
Solution Approach 2:
The machine learning models automatically adjust and refine their predictions based on feedback from actual system performance and workload patterns. The system self-optimizes by continuously learning from historical data and environmental factors, eliminating the need for manual tuning of complex assignment parameters. This self-service capability allows the system to handle dynamic workload changes effectively while maintaining relatively simple operational procedures, as the complexity is absorbed by the autonomous learning process rather than manual intervention.
3Reliability
If machine learning models are used to predict workload changes and optimize policy assignment, then load balancing and performance improve, but computational resources and processing time increase
Solution Approach 1:
The machine learning models are trained in advance using historical workload data, environmental event data, and performance metrics from multiple computing clusters. This offline training phase allows the models to learn complex patterns and relationships without impacting real-time assignment performance. When actual workload assignment is needed, the pre-trained models can quickly predict optimal cluster assignments based on current conditions, providing reliable load balancing decisions with minimal processing delay. The time-consuming training occurs beforehand, not during critical assignment operations.
Solution Approach 2:
The system employs multiple ML models with different levels of complexity and prediction horizons. For routine workload assignments, simpler and faster models are used that provide sufficient accuracy with minimal processing time. For more complex or critical assignments, the system can invoke more comprehensive models that analyze additional factors and provide more reliable predictions. This tiered approach ensures that the system achieves high reliability where needed while minimizing processing time for standard operations, balancing accuracy requirements with computational efficiency.
Data Source
AI summary
Presented herein are systems and methods of assigning policies across application instances using machine learning (ML) models. A computing system of a policy administration system may identify a first data structure of a first policy. The computing system may obtain a first plurality of attributes associated with the first policy. The computing system may apply a ML model to the first data structure and the first plurality of attributes. The ML model may be trained using a plurality of instance assignments. The computing system may assign, from applying the ML model, the first data structure of the first policy to a first application instance from the plurality of application instances of the policy administration system.