Virtualized Storage Mapping Trees for Address Management

The patent segments the storage system into distinct layers: a control layer handling logical address mapping and a data layer handling physical storage operations. This segmentation allows independent optimization of each layer, reducing overall system complexity while maintaining management efficiency. The control layer uses mapping trees to organize address mappings in a hierarchical structure, making large-scale address mapping manageable through divided responsibility.

The patent introduces a hierarchical dimension to address mapping by implementing mapping trees with multiple levels. Instead of flat one-to-one mapping, the system uses multi-level tree structures where internal nodes represent intermediate address ranges and leaf nodes represent actual mappings. This dimensional transformation enables efficient management of large address spaces by organizing mappings hierarchically rather than linearly.

The patent implements preliminary action by pre-allocating physical storage blocks and creating placeholder entries in the mapping trees before actual data migration occurs. When physical addresses change, the system has already prepared the mapping structure, reducing the need for frequent structural updates. The mapping trees are designed to accommodate address changes through pre-established node structures that can be efficiently updated without reconstructing the entire tree.

The patent uses copying mechanisms where changes to physical address mappings are first applied to copy or backup mapping structures, then validated and committed to the main mapping trees. This copying approach allows the system to maintain accurate mappings during transitions while avoiding performance degradation from direct in-place updates. The copy-then-commit strategy ensures mapping accuracy while optimizing update performance.