Distributed Storage Client Direct Node Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed storage systems, clients lack information about partition distribution and storage node locations, leading to high resource consumption and bandwidth overhead for access nodes when assembling data from multiple nodes.
Innovation Solution
An apparatus and method that provide clients with information about the running status and metadata of the distributed storage system, enabling them to directly access storage nodes and assemble data partitions, thereby reducing the load on access nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If access nodes assemble data partitions from multiple storage nodes, then data access reliability is improved, but resource consumption and bandwidth overhead increase significantly
Solution Approach 1:
The patent segments the data access function into two parts: (1) the access node provides metadata and running status information to the client, and (2) the intelligent client directly accesses storage nodes to fetch data partitions. This segmentation eliminates the bottleneck where access nodes previously had to assemble all data partitions, reducing their resource consumption while maintaining reliability through distributed access.
Solution Approach 2:
The patent enables intelligent clients to perform self-service by providing them with partition information and storage location metadata. Clients can independently locate and access data partitions from storage nodes without requiring access nodes to assemble the data, thereby reducing the load on access nodes while maintaining data access reliability through direct client-to-storage-node connections.
2Ease of operation
If access nodes assemble data from multiple storage nodes, then data access functionality is improved, but bandwidth overhead increases significantly
Solution Approach 1:
The patent extracts the data transfer function from access nodes and transfers it directly to clients. Access nodes only provide metadata and running status information, while clients directly fetch data partitions from storage nodes. This extraction eliminates the intermediate data transfer through access nodes, significantly reducing bandwidth overhead while maintaining full data access functionality.
3Device complexity
If partition information is transparent to clients, then system simplicity is improved, but access node load increases
Solution Approach 1:
The patent applies partial transparency by providing clients with specific partition information and storage location metadata needed for direct access, rather than complete transparency of all system details. This partial information provision enables clients to perform direct access while access nodes maintain oversight of system state through collected running status information, balancing simplicity with reduced access node load.
Data Source
AI summary
Embodiments of the present disclosure relate to an apparatus and method for facilitating access to data in a distributed storage system by using a processing unit configured to collect information related to running status of the distributed storage system; obtain metadata related to data stored in the distributed storage system; in response to a request from a client, provide to the client the information related to running status and metadata related to data to be accessed; and extend storage nodes in the distributed storage system so that the client can directly access the storage nodes by using the information related to running status and the metadata related to data to be accessed. Other embodiments are also disclosed.