Mapped RAID Extent Shuffle for Neighborhood Matrix Balance

The patent applies local quality by creating neighborhood zones around each disk where a limited number of other disks are designated as neighbors. This local organizational structure simplifies management within each zone while distributing extents across multiple disks to maintain rebuild performance. The neighborhood matrix tracks these local relationships to enable efficient extent allocation and movement decisions.

The patent segments the disk pool into neighborhood zones where each disk has a defined set of neighbor disks. This segmentation allows extents to be distributed across disks while maintaining manageable local relationships. The neighborhood matrix divides the complex global distribution problem into smaller local neighborhood management tasks, reducing overall system complexity.

The patent changes the parameters of extent movement decisions by introducing a scoring mechanism based on neighborhood matrix values. When moving extents between disks, the system calculates scores based on how the movement affects neighborhood balance and overlap reduction. This parameter-based approach enables efficient management of hybrid disk capacities by making automated decisions that balance storage efficiency with neighborhood relationships.

The patent applies partial action by performing shuffle operations selectively rather than exhaustively. The system identifies specific extent movements that provide the greatest benefit in reducing overlap and balancing neighborhoods, then executes only those movements. The scoring mechanism ranks potential movements, and the system performs operations in order of benefit until diminishing returns are reached, avoiding unnecessary shuffle operations and reducing time loss.