Adaptive Data Replication in Dispersed Storage Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
RAID systems face issues with disk failures leading to data loss, increased maintenance costs, and security concerns due to multiple data copies, as well as vulnerability to natural disasters and power outages.
Innovation Solution
A dispersed storage network (DSN) utilizing error encoding and decoding with Cauchy Reed-Solomon encoding, where data is split into encoded slices stored across multiple geographically diverse sites, allowing for adaptive replication and secure storage without redundant copies, enabling data recovery even with multiple storage unit failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is copied to multiple RAID devices for redundancy, then data loss risk is reduced, but security risk increases due to multiple accessible copies
Solution Approach 1:
The patent segments data into multiple encoded slices distributed across different storage units. Unlike traditional RAID that creates redundant copies, this segmentation uses error correction encoding where any subset of slices can reconstruct the original data, providing both redundancy and enhanced security since no single slice contains complete information.
Solution Approach 2:
The system changes the parameter of data representation from identical copies to mathematically transformed encoded slices. Using Cauchy Reed-Solomon encoding, the data is transformed into multiple slices where each slice has different mathematical properties, allowing flexible reconstruction while maintaining security.
2Quantity of substance
If more disks are added to RAID array for capacity, then storage capacity increases, but probability of disk failure increases
Solution Approach 1:
The patent implements beforehand cushioning by incorporating error correction capability into the data encoding process. The system can withstand a predetermined number of disk failures without data loss, providing a buffer against failures that occurs naturally as more disks are added to the array.
3Reliability
If data is stored in dispersed locations for disaster recovery, then resilience to natural disasters improves, but data access performance may deteriorate
Solution Approach 1:
The patent applies partial action by allowing data reconstruction from a subset of encoded slices rather than requiring all slices. The system is designed to tolerate loss of some slices while still enabling data recovery, thus maintaining access performance even when some dispersed storage locations are unavailable.
Data Source
AI summary
Adaptive replication of data in a dispersed storage network (DSN) to improve data access performance. In various examples, a DSN storage unit determines that a frequency of slice access of an encoded data slice stored by the storage unit compares unfavorably to a first slice access threshold (e.g., a greater number of accesses than a threshold number of accesses over a given period of time). The storage unit then identifies at least one secondary storage unit and replicates the encoded data slice to generate a replicated encoded data slice. The replicated encoded data slice is then sent to the at least one secondary storage unit for storage therein. In addition, a slice storage location table is updated to associate the at least one secondary storage unit and the replicated encoded data slice such that future access requests for the encoded data slice may be re-directed to a secondary storage unit.