Automated Storage Tiering via Simulation Engine
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional auto-tiering systems face inefficiencies and performance issues due to time-consuming data relocation processes, requiring repetitive try-and-error methods to determine the best configuration, which affects system performance and efficiency.
Innovation Solution
An automated tiering system and method that uses algorithm analyzers, a simulation engine, and a data migrator to simulate data relocation results, generate exploitation maps based on usage frequency, and evaluate configurations to identify the best system configuration for data migration, reducing the time required for data relocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional repetitive try-and-error data relocation is performed to determine the best configuration, then the system can learn the performance raised with each configuration, but the data relocation process becomes time-consuming and affects system performance
Solution Approach 1:
The patent applies preliminary action by performing simulation before actual data relocation. The simulation engine predicts the performance results of different configurations without actually moving data, allowing the system to identify the best configuration in advance. This eliminates the need for repetitive try-and-error relocation processes that traditionally consumed hours or days, reducing data relocation time while maintaining accurate performance evaluation through simulated results
2Adaptability or versatility
If repetitive data relocation is performed for each configuration candidate, then the system can identify the best configuration, but system performance is affected during the relocation processes
Solution Approach 1:
The patent applies copying by creating virtual copies of data allocation maps and exploitation maps in the simulation environment. Instead of actually relocating data multiple times for each configuration candidate, the simulation engine creates copies of the data structure and performs virtual relocation operations. This allows comprehensive configuration evaluation while maintaining system productivity, as the actual storage system remains undisturbed during the evaluation process
3Ease of manufacture
If traditional data relocation methods are used without simulation, then the process is simpler to implement, but the relocation process cannot be predicted in advance and requires actual execution to learn results
Solution Approach 1:
The patent introduces an intermediary simulation engine that acts as a mediator between the configuration evaluation requirement and the actual data relocation process. The simulation engine receives configuration parameters, performs virtual data relocation using algorithm analyzers, and predicts performance outcomes without affecting the actual storage system. This intermediary layer maintains implementation simplicity while dramatically reducing the time required to determine the best configuration by providing advance predictions
Data Source
AI summary
An automated tiering system and an automated tiering method are provided. The system includes a controller and multiple storage apparatuses that are layered into at least two tiers according to performance. In the method, an algorithm analyzer corresponding to each of multiple system configurations is executed to analyze data blocks in each storage apparatus to determine a target block of each data block after relocation and generate an estimated data allocation map. Then, a simulation engine is executed to classify the target blocks in the data allocation map according to a usage frequency of each target block so as to generate an exploitation map, and evaluate all of the exploitation maps to find the system configuration that raises the most performance as a best configuration. Finally, a data migrator is executed to migrate the data blocks in the storage apparatus according to the best configuration.