Skip to content

Latest commit

 

History

History
114 lines (81 loc) · 3.31 KB

README.md

File metadata and controls

114 lines (81 loc) · 3.31 KB

Distributed File System

A simple yet functional distributed file system implementation inspired by a CMU course project. The system features a centralized Naming Server architecture with distributed Storage Servers, supporting file replication and lock-based consistency.

🏗️ Architecture

Naming Server

  • Acts as the central coordinator between clients and storage servers
  • Manages file locations and lock distribution
  • Single instance running at any time
  • Handles server registration and file directory updates

Storage Servers

  • Multiple instances can run simultaneously
  • Register with Naming Server on startup
  • Maintain local file directories
  • Support file replication and consistency mechanisms

🔐 Lock Management

The system implements two types of locks for file access:

  • Shared Lock: Required for read operations
  • Exclusive Lock: Required for write operations

Note: Due to architectural constraints, shared locks are treated as write operations, and exclusive locks are treated as read operations at the implementation level.

📦 Features

Automatic Replication

  • New replicas are created after every 20 shared locks on a file
  • Replicas are distributed across available servers
  • Ensures non-blocking access for multiple clients

Consistency Management

  • Exclusive lock requests trigger replica consolidation
  • Maintains single source of truth for write operations
  • Automatic conflict resolution during server registration

🚀 Getting Started

Prerequisites

  • Go runtime environment
  • Available network ports for services
  • Storage space for file systems

Starting the Naming Server

Navigate to the naming directory and run:

go run . <Client_Service_Port> <Slave_Registration_Port>

Example:

go run . 8080 8081

Starting a Storage Server

Navigate to the storage directory and run:

go run . <Client_Service_Port> <Command_Port> <Registration_Port> <Storage_Directory> <Naming server ip>

Example:

go run . 9090 9091 8081 storage1 localhost

💡 Implementation Notes

Storage Directory Handling

  • Automatically created if it doesn't exist
  • Preserves existing contents if directory is present
  • Handles file conflicts during registration

Client Expectations

  1. Clients must request appropriate locks before file operations
  2. Locks should be released after operations complete
  3. Cooperative access model is assumed

🔍 System Assumptions

  1. Clients follow the locking protocol:

    • Acquire appropriate locks before operations
    • Release locks after completing operations
    • Use shared locks for reads and exclusive locks for writes

  2. File Replication:

    • System maintains consistency across replicas
    • Latest version access is guaranteed
    • Non-blocking operations during normal usage

📚 Technical Details

Replication Mechanism

  • Triggered after 20 shared locks
  • Distributed across available servers
  • Automatic cleanup during exclusive access

Consistency Protocol

  • Single-writer, multiple-reader model
  • Automatic conflict resolution
  • Replica consolidation during writes

🤝 Contributing

Feel free to submit issues and enhancement requests.

🙏 Acknowledgments

This project is inspired by a CMU course project. For detailed project requirements, refer to the original course materials.