Tiered Storage Policy Engine for SSD HDD Data Placement

The patent segments files into multiple parts and stores different parts in different storage media. Specifically, it divides files into segments that can be stored in SSD, HDD, or cloud storage based on access frequency and importance, allowing the system to optimize between performance and cost by placing only critical data in expensive SSD while keeping less critical data in cheaper storage.

The patent applies local quality by assigning different storage qualities to different parts of the same file based on their access patterns. Frequently accessed or critical portions of files are stored in high-performance SSD, while less frequently accessed portions are stored in cheaper HDD or cloud storage, thereby optimizing overall system performance without unnecessarily increasing cost for all data.

The system segments write operations and file data to utilize appropriate storage media. Critical and frequently accessed data is written to SSD for high performance, while less critical data is written to HDD or cloud storage, allowing the system to maintain good write performance for important data while keeping overall storage costs low.

The patent applies local quality by assigning different write performance characteristics to different parts of the file system. Important directories and files receive high-performance SSD storage with fast write capabilities, while other data is stored in cheaper HDD or cloud storage with acceptable but slower write performance, thereby optimizing the balance between cost and productivity.

Instead of migrating entire volumes to SSD, the patent segments volumes into smaller units that can be individually evaluated and placed in appropriate storage media based on access patterns. This allows only the necessary portions of data to be stored in expensive SSD, rather than the entire volume, thereby reducing cost while maintaining performance benefits for critical data.

The patent applies local quality by assigning different storage qualities to different portions of the volume based on local access patterns rather than treating the entire volume uniformly. This allows high-performance SSD storage to be applied only to locally hot data while cooler data remains in cheaper storage, optimizing the balance between read performance and cost.

The patent introduces an intermediary storage management layer that handles the complexity of fine-grained data placement automatically. This intermediary layer includes components like the storage manager, policy engine, and data movement services that transparently manage data segmentation, placement, and migration between different storage media based on policies, thereby enabling storage optimization without requiring complex manual intervention.

The system implements self-service capabilities where the storage management infrastructure automatically monitors access patterns, evaluates data importance, and performs data placement and migration decisions without manual intervention. The policy-based automation and self-managing features reduce operational complexity while maintaining fine-grained storage optimization.