Greedy Packing Algorithm for Storage Track Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-capacity data storage systems like SANs face inefficiencies in resource utilization due to fixed-size back-end tracks, leading to wasted space when packing variable-size front-end tracks, which affects metadata management and drive utilization.
Innovation Solution
Implement a method where compute nodes in a data storage system sort front-end tracks by size, match them with cached packing solutions, and pack them into slices, using a greedy first fit packing algorithm to optimize space utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If fixed-size back-end tracks are used for storage allocation, then metadata management becomes simpler and more practical, but storage resource utilization efficiency decreases due to unused space in tracks
Solution Approach 1:
The patent segments front-end tracks of varying sizes into groups and packs multiple smaller tracks into single back-end tracks. This segmentation allows efficient utilization of fixed-size back-end tracks while maintaining simplified metadata management, as the system only needs to track packed units rather than individual variable-size tracks.
Solution Approach 2:
The patent merges multiple front-end tracks into single back-end tracks through a packing mechanism. By combining several smaller tracks into one larger fixed-size track, the system eliminates unused space and improves storage utilization while keeping metadata management straightforward.
2Loss of energy
If variable-size front-end tracks are packed into fixed-size back-end tracks, then storage resource utilization improves, but the complexity of the packing algorithm and processing increases
Solution Approach 1:
The patent applies preliminary actions by pre-sorting front-end tracks into buckets based on size categories before packing. This pre-processing step simplifies the subsequent packing algorithm, as tracks are already organized by size, making it easier to match them with appropriate back-end tracks and reducing computational complexity.
Solution Approach 2:
The patent changes the parameter of track organization by sorting tracks into size-based buckets. This parameter transformation converts the packing problem from handling arbitrary variable-size tracks to systematically matching tracks within size categories, simplifying the packing algorithm while maintaining high storage utilization.
3Productivity
If larger capacity back-end tracks are used, then memory and processing burden for metadata management is reduced, but drive utilization efficiency decreases due to more unused space
Solution Approach 1:
The patent introduces a new dimension of organization by implementing size-based bucketing of front-end tracks. This additional organizational dimension allows the system to efficiently pack tracks of various sizes into fixed-size back-end tracks, simultaneously achieving good metadata management scalability and high drive utilization.
Solution Approach 2:
The patent creates a mapping structure that copies the essential information about packed tracks into metadata. This copying approach allows the system to manage large back-end tracks efficiently by maintaining simplified metadata representations that track the contents and usage of each packed unit without requiring complex management of every individual track.
Data Source
AI summary
A storage array packs multiple non-full-size front-end tracks into slices that contain multiple back-end tracks. A greedy first fit packing algorithm is used to find packing solutions that are cached and ranked. The cached, ranked packing solutions are used by attempting to find matches with bucketed front-end tracks to be relocated. New packing solutions are generated and cached when matches cannot be found. Packing solutions may be shared outside the domain in which they are discovered.