Storage System Reliability-Based Data Allocation and Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network storage technologies face performance and reliability issues due to the lack of semantic understanding of data at the physical layer, leading to inefficient data operations and increased vulnerability during recovery processes, as magnetic storage media density grows without corresponding speed improvements.
Innovation Solution
A storage system that assigns reliability values to logical containers based on Service Level Objectives (SLOs), allowing for prioritization of data operations by identifying and utilizing specific parity groups with varying protection levels, thereby enabling the physical layer to distinguish and prioritize data storage and recovery processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If magnetic storage media density is increased, then storage capacity is improved, but data operation speed deteriorates
Solution Approach 1:
The patent divides the storage system into multiple parity groups with different protection levels (e.g., RAID 5, RAID 6, RAID 10) assigned to different data regions. High-reliability data receives enhanced protection with faster recovery capabilities, while less critical data uses standard protection. This local differentiation allows the system to optimize recovery speed for critical data without sacrificing the high density benefits across the entire storage array.
2Reliability
If data recovery operations are performed on all parity groups equally, then complete data protection is achieved, but system performance during recovery deteriorates
Solution Approach 1:
The system pre-classifies data into different reliability tiers and assigns appropriate parity groups before failures occur. When a failure happens, the system immediately identifies which parity group is affected and prioritizes recovery based on the pre-assigned reliability level. This preliminary classification eliminates the need for complex real-time decision-making during recovery, maintaining high system performance while ensuring critical data is restored first.
Solution Approach 2:
The storage system is segmented into multiple parity groups with different protection levels. Instead of treating all data uniformly, the system divides data storage into distinct segments (parity groups) that can be independently managed and recovered. This segmentation allows parallel recovery operations across different parity groups and enables the system to focus resources on critical data segments first, maintaining overall system productivity during recovery operations.
3Device complexity
If uniform data layout is used across all storage devices, then system simplicity is maintained, but data operation efficiency deteriorates
Solution Approach 1:
The patent implements a hierarchical data layout where different regions of the storage array are assigned different parity group types based on data reliability requirements. The system maintains simplicity at the interface level (uniform access methods) while introducing complexity at the internal allocation level (different parity groups for different data types). This allows efficient data operations for critical information without significantly increasing overall system complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A storage system provides highly flexible data layouts that can be tailored based on reliability considerations. The system allocates reliability values to logical containers at an upper logical level of the system based, for example, on objectives established by reliability SLOs. Based on the reliability value, the system identifies a specific parity group from a lower physical storage level of the system for storing data corresponding to the logical container. After selecting a parity group, the system allocates the data to physical storage blocks within the parity group. In embodiments, the system attaches the reliability value information to the parity group and the physical storage units storing the data. In this manner, the underlying physical layer has a semantic understanding of reliability considerations related to the data stored at the logical level. Based on this semantic understanding, the system has the capability to prioritize data operations on the physical storage units according to the reliability values attached to the parity groups.