Single-Node Server Data Storage via Erasure Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional single-node server data storage systems rely on expensive RAID hardware to achieve disaster recovery, making it difficult to balance hardware resource costs and disaster recovery capabilities.
Innovation Solution
A data storage method that processes original data into N first data blocks and M second data blocks, with an association relationship, and stores them across different disks of a single-node server, allowing for flexible disaster recovery without relying on hardware RAID solutions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional RAID hardware is used to provide disaster recovery capability, then disaster recovery capability is improved, but hardware resource cost increases
Solution Approach 1:
The patent replaces the mechanical RAID hardware system with a software-based erasure coding system. Instead of using dedicated RAID controllers and hardware, the invention implements disaster recovery through software algorithms that perform data slicing and encoding processing on standard disk drives, thereby eliminating the need for expensive RAID hardware while maintaining disaster recovery capabilities.
Solution Approach 2:
The patent changes the approach from hardware-based RAID levels (RAID5, RAID6) to a software-based erasure coding system with configurable parameters N and M. This allows flexible adjustment of data block distribution and redundancy levels through software parameters rather than fixed hardware configurations, enabling cost-effective disaster recovery without sacrificing reliability.
2Reliability
If RAID hardware is deployed to ensure data protection, then data reliability is improved, but storage space efficiency deteriorates
Solution Approach 1:
The patent introduces configurable parameters N and M to control the balance between data protection and storage efficiency. By adjusting these parameters, the system can optimize the ratio of data blocks to redundant blocks, allowing flexible control over storage space utilization while ensuring adequate data protection without the overhead of traditional RAID hardware.
Solution Approach 2:
Instead of using traditional RAID duplication methods that copy entire data blocks, the patent employs erasure coding to create distributed data blocks with mathematical redundancy. This allows more efficient use of storage space by creating smaller, distributed redundant units that can be reconstructed from fewer original blocks, improving storage efficiency while maintaining data protection.
3Reliability
If hardware RAID is used to provide multi-level disaster recovery, then disaster recovery capability is improved, but hardware resource cost increases
Solution Approach 1:
The patent replaces the mechanical RAID hardware system with a software-based erasure coding system. Instead of using dedicated RAID controllers and hardware, the invention implements disaster recovery through software algorithms that perform data slicing and encoding processing on standard disk drives, thereby eliminating the need for expensive RAID hardware while maintaining disaster recovery capabilities.
Solution Approach 2:
The patent creates a universal storage system where any standard disk drive can function as part of the RAID array without requiring specialized hardware. The software-based approach allows the same hardware to provide multiple functions including data storage, error correction, and disaster recovery, eliminating the need for separate RAID hardware and reducing overall system cost.
Data Source
AI summary
The present disclosure provides a data storage method, a single-node server, and a device. The method includes: obtaining storage parameters of at least one single-node server, wherein the storage parameters represent parameters used when slicing processing and encoding processing are performed on original data to be stored; according to the storage parameters of the single-node server, determining N first data blocks and M second data blocks which are obtained after said original data is processed, wherein any N data blocks in the N first data blocks and the M second data blocks have an association relationship with M data blocks other than the any N data blocks in the N first data blocks and the M second data blocks; and respectively storing data blocks corresponding to the original data into different disks of the single-node server.