RAID Drive Utilization via Vertical Log-Structured Writing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing RAID systems face challenges in efficiently managing drive utilization, leading to increased computing power, processing power, short-term memory usage, bandwidth issues, manual labor, downtime, financial loss, and reduced storage efficiency, especially in the event of multiple drive failures.
Innovation Solution
A system and method that directs log-structured writing of data vertically across chunks of a single physical drive for each logical volume, while also writing parity data across multiple physical drives, thereby optimizing drive utilization and reducing the impact of drive failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is written horizontally across multiple physical drives in traditional RAID systems, then storage redundancy is achieved, but drive utilization efficiency deteriorates and computing power consumption increases
Solution Approach 1:
The patent segments data writing operations into two distinct patterns: vertical writing for data chunks within a single physical drive, and horizontal writing for parity data across multiple drives. This segmentation allows optimization of each writing pattern independently, improving overall drive utilization while maintaining redundancy through the horizontal parity distribution.
Solution Approach 2:
The patent introduces a vertical dimension to traditional horizontal RAID writing by implementing vertical striping within individual drives before horizontal parity distribution. This dimensional change transforms the writing pattern from purely horizontal across drives to a combination of vertical within drives and horizontal across drives, enhancing utilization efficiency.
2Productivity
If traditional horizontal striping is used across multiple drives, then data distribution is achieved, but processing power and bandwidth consumption increase
Solution Approach 1:
The patent segments the data distribution process into vertical chunk writing within single drives and horizontal parity writing across drives. This reduces processing overhead by localizing vertical writes to single drives, minimizing cross-drive bandwidth consumption for data chunks, while parity calculations are performed separately.
Solution Approach 2:
The patent introduces log-structured virtual volumes as an intermediary layer between the host system and physical drives. This intermediary abstracts the complex horizontal striping operations, presenting a simplified vertical writing interface to hosts while managing the actual horizontal parity distribution in the background, thereby reducing processing complexity.
3Reliability
If excessive backup storage is allocated for multiple drive failures, then data protection is improved, but storage efficiency deteriorates
Solution Approach 1:
The patent implements partial redundancy by writing parity data horizontally across a subset of drives rather than all drives. This partial action provides sufficient protection against single drive failures while avoiding the excessive storage overhead of traditional RAID configurations that protect against multiple simultaneous failures, thereby improving storage efficiency.
Solution Approach 2:
The patent changes the redundancy parameter from traditional horizontal striping across all drives to vertical striping within drives combined with selective horizontal parity distribution. This parameter change reduces the amount of storage dedicated to redundancy while maintaining adequate protection, improving overall storage efficiency.
Data Source
AI summary
One or more systems, devices, computer program products, and/or computer-implemented methods provided herein to use a redundant array of disks. A system can comprise a memory that stores computer executable components, and a processor that executes the computer executable components stored in the memory, wherein the computer executable components can comprise a control component that directs, for n physical drives of a redundant array of disks (RAID) storing data for at least n logical volumes, log-structured writing of data of each logical volume of the at least n logical volumes vertically across chunks of only a single physical drive of the n physical drives, wherein the control component further directs writing of parity data at each of the physical drives, which parity data at each physical drive of the subset respectively corresponds to other ones of the physical drives of the n physical drives.