Virtual Disk Mapping via Hierarchical Sparse Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Storage virtualization systems face issues with data fragmentation, performance degradation, and high metadata overhead due to scattered allocation of physical blocks in thin provisioning methods, leading to inefficient storage utilization and increased administrative costs.
Innovation Solution
The implementation of an intelligent storage application resource (iSAR) within a Fibre Channel fabric that uses hierarchical sparse mapping and deduplication to allocate logical blocks to physical disks, reducing fragmentation and metadata overhead by ensuring data blocks are stored in close proximity and compressing data to optimize storage space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If thin provisioning is used to allocate physical blocks dynamically, then storage utilization is improved, but data fragmentation increases
Solution Approach 1:
The patent segments the storage space into units and subunits, creating a hierarchical structure. This segmentation allows for organized allocation of data blocks while maintaining track of their locations, thereby reducing fragmentation issues while preserving storage utilization benefits.
Solution Approach 2:
The patent introduces a mapping structure as an intermediary between virtual storage and physical storage. This mapping tracks the location of data blocks and enables efficient management of fragmented data, allowing the system to maintain high storage utilization while mitigating the performance impact of fragmentation.
2Quantity of substance
If physical blocks are scattered over entire physical storage pool, then storage capacity is improved, but I/O operation performance degrades
Solution Approach 1:
By segmenting storage into hierarchical units and subunits, the patent creates manageable sections that can be efficiently accessed. This segmentation enables the system to work with smaller, more localized data sets, improving I/O performance while maintaining overall storage capacity.
Solution Approach 2:
The patent performs preliminary organization of data blocks into units and subunits with associated mapping information. This preliminary structuring enables faster access to data during I/O operations, as the system doesn't need to search the entire storage pool but can directly access pre-organized segments.
3Productivity
If thin provisioning is implemented, then storage allocation efficiency is improved, but metadata overhead increases
Solution Approach 1:
The hierarchical segmentation of storage into units and subunits creates a more compact metadata structure. Instead of tracking every individual block across the entire storage pool, the system only needs to maintain mapping information at the unit and subunit levels, significantly reducing metadata overhead while preserving allocation efficiency.
Data Source
AI summary
A storage area network can include a storage virtualization entity—intelligent storage application resource (iSAR)—either as a separate device in the fabric, or as an integrated module in one or more switches within the fabric. All I/O operations can be re-directed to iSAR for processing. iSAR can segment virtual storage and physical storage into units, where each unit of the virtual storage is mapped to a single unit in physical storage. Data associated with incoming I/O operation can be compressed before being stored in physical storage. iSAR includes overflow reserve storage at the block, sub-page and page level to accommodate changes in compressed data size on subsequent I/O operations. These measures can improve I/O performance and reduce fragmentation. iSAR can also employ deduplication of incoming data stored on physical storage to improve storage efficiency.