Cloud Storage Tiering via ML Temperature Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud providers face challenges in optimizing the allocation of storage objects across different tiers in cloud storage systems to maximize performance while minimizing costs, due to varying access patterns and cost-access tradeoffs.
Innovation Solution
A computer-implemented method that forecasts a future temperature for storage objects in a multi-tiered cloud storage system using time-series models or unsupervised machine learning models, determining the cost of modifying tiering based on this forecast, and adjusting tiering policies accordingly to optimize storage object placement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If storage objects are placed in higher performance cloud storage tiers, then access time is reduced, but storage cost increases
Solution Approach 1:
The system dynamically adjusts storage tier allocation based on forecasted future temperatures (access patterns). Instead of static tiering, the system continuously updates temperature forecasts using machine learning models and re-positions storage objects between tiers to optimize the trade-off between access speed and storage cost, allowing the configuration to adapt over time
Solution Approach 2:
The system performs preliminary forecasting of future access patterns using machine learning models before making tiering decisions. By predicting future temperature values, the system proactively positions storage objects in appropriate tiers in advance, rather than reactively responding to past access patterns, thereby optimizing both access time and cost
2Loss of energy
If storage objects are placed in lower performance cloud storage tiers, then storage cost is reduced, but access time increases
Solution Approach 1:
The system dynamically adjusts storage tier allocation based on forecasted future temperatures. By continuously monitoring predicted access patterns and adjusting tier positions accordingly, the system ensures that objects requiring frequent access remain in higher performance tiers while minimizing storage costs for less accessed objects
Solution Approach 2:
The system changes the temperature parameter (access pattern metric) used for tiering decisions from historical data to forecasted future data. This parameter change enables the system to make forward-looking tiering decisions that optimize both cost and access time based on predicted rather than past behavior
3Device complexity
If simplified reactive tiering mechanisms are used, then device complexity is reduced, but productivity decreases
Solution Approach 1:
The system implements self-service through automated machine learning models that independently forecast temperature values and determine optimal tiering decisions without manual intervention. The system self-adjusts storage object positions based on predicted access patterns, eliminating the need for complex manual tiering management while maximizing optimization productivity
Solution Approach 2:
The system incorporates feedback loops where actual access patterns are compared with forecasted temperatures, and the machine learning models are retrained and refined based on this feedback. This continuous improvement mechanism enhances prediction accuracy and tiering optimization over time without increasing operational complexity
Data Source
AI summary
A method, computer program product, and computing system for forecasting a future temperature for a storage object within a multi-tiered cloud storage system. A cost associated with modifying a tiering of the storage object within the multi-tiered cloud storage system is determined based upon, at least in part, the future temperature forecasted for the storage object. The storage object is tiered in the multi-tiered cloud storage system based upon, at least in part, the cost associated with modifying the tiering of the storage object and a tiering policy associated with the multi-tiered cloud storage system.