Shared Parity RAID Storage System for SSD HDD Hybrid Workloads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional RAID systems face performance bottlenecks due to inefficient use of SSDs and increased failure rates in enterprise data centers, where SSDs are not fully utilizing their performance characteristics and have smaller storage capacity and higher costs compared to HDDs, and parity access becomes a bottleneck without additional hardware resources.
Innovation Solution
A data storage system where multiple RAID volumes share a single set of parity disks, allowing parity information to be written in parallel or sequentially, reducing interference and enhancing write performance, and using an inverted list for sequential reads and writes during recovery to improve I/O performance and reduce data recovery impact on applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional RAID systems use SSDs for data storage, then random read speed and IOPS are improved, but storage capacity is reduced and price increases
Solution Approach 1:
The storage system is segmented into two distinct layers: SSD layer for hot data requiring high-speed access, and HDD layer for cold data requiring high capacity. This segmentation allows each storage type to be optimized for its specific function, resolving the contradiction between speed and capacity
Solution Approach 2:
The patent introduces a new dimension of data classification based on access frequency (hot vs. cold data), organizing storage resources across multiple dimensions rather than using a single storage type for all data. This dimensional approach enables simultaneous optimization for both speed and capacity
2Productivity
If RAID algorithms are applied to SSDs, then IOPS performance is improved, but erase-before-write and wear-out issues increase
Solution Approach 1:
The patent segments the storage workload by creating separate RAID volumes for SSDs and HDDs, with SSDs handling hot data that requires high IOPS. This segmentation protects SSDs from excessive write amplification and wear while maintaining high productivity for time-sensitive operations
Solution Approach 2:
The system dynamically changes operational parameters by adjusting which RAID volumes are active based on workload characteristics. When workload is write-intensive, the system can reduce SSD usage or switch to HDD-based RAID volumes, changing the operational state to preserve SSD lifetime while maintaining productivity
3Reliability
If multiple RAID volumes are configured independently, then fault isolation is improved, but system-wide optimization is reduced
Solution Approach 1:
The patent implements a unified storage controller that manages multiple RAID volumes with different characteristics (SSD-based and HDD-based). This universal controller provides system-wide optimization through centralized resource allocation while maintaining fault isolation through independent RAID groups, resolving the contradiction between reliability and productivity
4Reliability
If parity information is stored in individual RAID volumes, then data protection is improved, but parity access becomes a performance bottleneck
Solution Approach 1:
The patent merges parity storage resources across multiple RAID volumes into a shared pool. Parity information for different RAID volumes is distributed across shared HDD resources, allowing parallel parity operations and eliminating the bottleneck of sequential parity access while maintaining data protection through distributed redundancy
Data Source
AI summary
The data storage system and the data storage approach based on redundant arrays of independent disks (RAID) is provided. The storage system includes: an RAID controller, and a storage device subsystem. This storage device subsystem includes a plurality of first physical storage devices for data storage and a plurality of second physical storage devices for parity storage. First physical storage devices are divided into a plurality of sets of first disks which are independent from each other. The data accesses to different sets of first disks will not interfere with each other. These second disks are shared by all the sets of first disks. The RAID controller stores data in each set of first disks and stores parity information in multiple shared second disks by the RAID mechanism. The storage system combines the parity from different RAID volumes into one shared set and this improves the performance and reliability of the whole storage system.