Storage Group Response-Time Prediction for SLO Breach Alerts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed storage systems face challenges in predicting and managing service level objectives (SLO) for storage groups, leading to potential breaches and inefficiencies in resource allocation and performance.
Innovation Solution
A machine learning model, such as a neural network, is used to classify and predict response times for storage groups, enabling proactive detection of SLO breaches by analyzing historical response time data and generating alerts or corrective actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional monitoring methods are used to detect SLO breaches, then the system can identify breaches after they occur, but the response time is delayed and corrective actions are taken too late
Solution Approach 1:
The machine learning model performs preliminary classification of response time data to predict future SLO breaches before they actually occur. By analyzing historical response time patterns and predicting future instances, the system takes preliminary action to alert operators in advance, allowing them to prepare corrective measures before the breach happens, thus eliminating the time loss associated with reactive monitoring.
2Loss of information
If manual analysis of response time data is performed, then the system can identify patterns, but the process is time-consuming and requires significant human resources
Solution Approach 1:
The patent replaces the mechanical process of manual data analysis with an automated machine learning system. The ML model automatically classifies response time data, identifies patterns, and predicts future breaches without human intervention. This substitution dramatically increases productivity by processing large volumes of data instantaneously while maintaining or improving the quality of pattern recognition compared to manual analysis.
3Device complexity
If the storage system operates without prediction capabilities, then the system structure remains simple, but the system cannot proactively manage resources or prevent performance degradation
Solution Approach 1:
The patent introduces a machine learning model as an intermediary component between the storage system's response time data and the SLO management process. This intermediary automatically processes raw response time data, classifies it according to SLO parameters, and generates predictions about future breaches. The intermediary adds minimal complexity to the system architecture while dramatically improving proactive SLO management capability, as it bridges the gap between raw data and actionable insights without requiring complex manual analysis or system redesign.
Data Source
AI summary
A method is provided, comprising: obtaining a plurality of measured response times for a given storage group, each of the measured response times corresponding to a different instance of a same time window; classifying the plurality of measured response times with a machine learning model to obtain a first predicted response time and a second predicted response time, the first predicted response time corresponding to a first future instance of the time window, and the second predicted response time corresponding to a second future instance of the time window; detecting whether the first predicted response time and the second predicted response time satisfies a service level objective for the given storage group; and generating a notification when at least one of the first predicted response time and the second predicted response time fails to satisfy the service level objective.