Distributed Storage Node Access State Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional distributed storage systems face performance bottlenecks and reliability issues due to the need to read all strips of a stripe during data access, leading to increased memory load and reduced system performance, especially in scenarios like random read operations.
Innovation Solution
The introduction of a state information set to record and manage access states of memories in the distributed storage system, allowing direct data read from trusted access states without performing complex correctness checks, and using this information to optimize data read and write operations by determining the validity of memory states and routing requests accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all data strips and parity strips in a stripe are read to perform data correctness check, then data reliability is ensured, but memory load increases and system performance is reduced
Solution Approach 1:
The patent applies partial action by reading only the necessary data strips containing the target data rather than all strips in the stripe. The system determines which specific storage nodes hold the required data strips based on strip distribution information, and reads only those specific strips needed to reconstruct the target data, avoiding the excessive reading of all N data strips and M parity strips that would be required for a full correctness check.
2Quantity of substance
If stripe length is increased to store more data, then storage capacity is improved, but load of hard disk increases and access efficiency is reduced
Solution Approach 1:
The patent segments the large stripe into individual data strips that are distributed across different storage nodes. When accessing target data, the system identifies and accesses only the specific segmented strips containing the target data rather than reading the entire large stripe. This segmentation allows the system to maintain large storage capacity while improving access efficiency by reading only the necessary portions.
3Device complexity
If centralized storage server is used to store all data, then system structure is simplified, but system performance bottleneck and reliability weakness occur
Solution Approach 1:
The patent segments the centralized storage architecture into multiple distributed storage nodes, each holding portions of the data strips. This segmentation eliminates the single-point performance bottleneck of centralized storage while maintaining manageable system complexity through standardized node interfaces and centralized management of strip distribution information.
Solution Approach 2:
The patent introduces a management component that acts as an intermediary, maintaining centralized information about strip distribution across storage nodes. This intermediary enables the system to achieve distributed storage performance while preserving simplified access control and data location management, bridging the gap between distributed architecture and centralized management simplicity.
Data Source
AI summary
In a distributed storage system, a first storage node receives a first data read request for reading first data. The first data request includes a first address of the first data. The first storage node determines that the first data are stored in a first storage device included in the first storage node based on the first address. Based on the determination, the first storage node checks whether the first storage device is in a trusted access state. Further, the first storage node obtains the first data when the first storage device is in the trusted access state.