Dynamic Event Prioritization in Networked Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current networked storage systems rely on static severity levels to prioritize event resolution, which are inflexible and do not adapt to varying operating environments, leading to inefficient event handling and potential misallocation of resources.
Innovation Solution
Implement a machine learning-based system that assigns dynamic priority scores to events using a training dataset and iterative predictive algorithms, considering multiple parameters such as event source, impact area, and resource utilization, to automate corrective actions and optimize resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If static severity levels are used to prioritize events, then event handling follows a simple and consistent rule-based approach, but the system lacks adaptability to different operating environments and event contexts
Solution Approach 1:
The patent transforms the static severity level system into a dynamic one by introducing a machine learning model that continuously learns from historical event data and operational context. The prioritization scores are no longer fixed but adapt over time based on patterns learned from the training dataset, allowing the system to respond differently to the same event type under different operating conditions.
Solution Approach 2:
The invention changes the parameters used for event prioritization from simple static severity levels to multiple dynamic parameters including event type, source, impacted resources, and contextual factors. The machine learning model processes these varied parameters to generate prioritization scores, enabling the system to consider multiple dimensions of event importance simultaneously.
2Productivity
If multiple events with same severity levels are handled, then all events receive equal attention, but storage administrators become overwhelmed and cannot efficiently allocate resources
Solution Approach 1:
The system implements feedback by continuously monitoring event outcomes and using this information to refine prioritization scores. Historical event data including resolution times, impact severity, and operational context are fed back into the machine learning model to improve future prioritization accuracy, creating a self-improving system that learns from past performance.
Solution Approach 2:
The patent applies preliminary action by pre-calculating and storing prioritization scores for historical events during the training phase. This allows the system to quickly retrieve and apply learned prioritization patterns to new events without performing complex calculations in real-time, reducing the time required for event resolution.
3Reliability
If static severity levels are used for all events, then the system is simple to implement and maintain, but critical events may be misprioritized based on contextual factors like time of day or resource importance
Solution Approach 1:
The system applies self-service by enabling the machine learning model to automatically learn and adjust prioritization criteria without requiring manual configuration or intervention. The model trains itself on historical event data and operational context, autonomously identifying patterns and relationships that determine event importance, thereby improving reliability without proportional increases in operational complexity.
Solution Approach 2:
The patent uses preliminary action by pre-training the machine learning model with extensive historical event data before deployment. This offline training phase allows the system to learn complex prioritization patterns in advance, so that during operational use, the model can quickly and accurately prioritize events based on learned patterns without requiring complex real-time computations.
Data Source
AI summary
Methods and systems for a networked storage system are provided. One method includes utilizing a training dataset for prioritizing a plurality of events associated with a networked storage system using a plurality of resources. Each event is associated with a plurality of parameters, each parameter associated with a severity level determination for each event; and each event is provided an initial priority score based on a time when each event is selected for resolution. The plurality of parameters may include an event source. The method further includes using the training dataset to identify a weight of each parameter by executing an iterative prediction algorithm; determining a priority score of a new event based on the weight of each parameter; updating the training dataset with the priority score of the new event; and adjusting a resource impacted by the new event, based on the priority score of the new event.