Storage Cache Duplication via Save Memory Area

The patent saves data blocks into a save area before duplication begins. When duplication is triggered, the system checks if the required data blocks are already in the save area (pre-saved), and if so, performs duplication from memory without disk access. This preliminary action of pre-saving blocks eliminates the need for disk access during duplication, resolving the contradiction between maintaining duplication capability and preserving host I/O performance.

The patent introduces a save area in memory as an additional layer between the cache memory and the disk. Instead of creating logs only on disk (one dimension), the system maintains data blocks in both cache memory and save area (adding a memory dimension). This allows backup operations to read from the memory-based save area rather than disk, reducing device complexity and load while maintaining backup capability.

Instead of allocating enormous memory capacity to hold all possible data blocks (excessive action), the patent selectively saves only the data blocks that are targets of upcoming duplication operations into the save area (partial action). This selective approach reduces memory capacity requirements to realistic levels while still achieving the goal of reducing disk access during duplication by having the required blocks pre-loaded in memory.