Distributed File System Data Access via Log-Based State Restoration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network-based shared file systems, such as NFS/CIFS, face performance and scalability bottlenecks due to the need for a coordination server to manage data access consistency, leading to delays and read-write conflicts.
Innovation Solution
A distributed file system architecture that allows computing nodes to send mounting requests to storage nodes, which provide storage location information for log segments and checkpoints, enabling computing nodes to restore the data state of logical volumes and perform data access processing without relying on a central coordination server.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a coordination server is deployed to manage data access consistency using a lock competition mechanism, then data consistency is ensured, but system scalability deteriorates due to the coordination server becoming a bottleneck
Solution Approach 1:
The patent extracts the coordination server component from the system architecture and replaces it with a distributed lock-free mechanism. Each computing node independently manages its own data access without requiring centralized coordination, thereby eliminating the scalability bottleneck while maintaining data consistency through version vectors and conflict-free replicated data types (CRDTs).
Solution Approach 2:
The patent segments the centralized coordination function into distributed autonomous decision-making at each computing node. Instead of one central authority managing all locks, each node independently determines data access permissions using local state information and version vectors, distributing the coordination burden across the entire system.
2Reliability
If a lock competition mechanism is used to coordinate read and write operations, then data consistency is maintained, but data access delay increases due to clients waiting for locks
Solution Approach 1:
The patent enables each computing node to serve its own data access requests without external coordination. Nodes use local version vectors and conflict resolution algorithms to autonomously determine whether to proceed with read or write operations, eliminating the need to wait for lock grants from a central server and significantly reducing access delays.
Solution Approach 2:
The patent pre-computes and caches version vectors and data state information at each computing node before actual data access occurs. This preliminary preparation allows nodes to immediately determine access permissions and execute operations without waiting for real-time coordination, reducing latency while maintaining consistency.
3Productivity
If multiple clients access the same file data simultaneously, then system utilization is improved, but read-write conflicts occur requiring lock competition
Solution Approach 1:
The patent changes the fundamental parameters of data access by replacing binary lock states (locked/unlocked) with continuous version vectors that track data modifications. This allows multiple clients to simultaneously access data with different version stamps, enabling parallel operations without conflicts while maintaining consistency through version comparison and conflict resolution algorithms.
Solution Approach 2:
The patent creates local copies of data and version vector information at each computing node. Instead of sharing a single locked resource, multiple nodes maintain independent copies that can be accessed simultaneously, with consistency restored through periodic synchronization and conflict resolution based on version vectors, thereby eliminating read-write conflicts while improving utilization.
Data Source
AI summary
A method is implemented by a computing node, and includes: sending a mounting request triggered for a target logical volume to a target storage node of multiple storage nodes included in a file system; receiving storage location information of a log segment and a checkpoint corresponding to the target logical volume sent by the target storage node; reading log metadata of the log segment and the checkpoint according to the storage location information to restore a data state of the target logical volume; and performing data access processing based on the data state of the target logical volume. When multiple computing nodes access a target logical volume, read and write consistency of the multiple computing nodes can be maintained based on log information of the target logical volume maintained by various storage nodes. Furthermore, reading and writing can be performed synchronously, thus avoiding read and write conflicts.