Chandi-Lamport Distributed Snapshot Algorithm

Overview

This project implements the Chandi-Lamport Algorithm for recording a consistent global snapshot in a distributed system. The algorithm is applied within a token-passing system, where processes communicate over TCP and record their state and channel messages during snapshots.

Key Features

• Distributed snapshot capturing using marker messages.
• Token passing among multiple processes.
• Configurable delays for token propagation and marker transmission.

Requirements

• Docker
• Docker Compose

Usage

1. Clone the repository and navigate to the implementation.

2. Build the Docker image:

   docker build -t prj2 .
   

3. Run the program using Docker Compose:

   docker-compose -f docker-compose-testcase-[1-5*].yml up
   

   This will start 5 peer containers as defined in the docker-compose-testcase-[1-5*].yml file.

4. To change the number of peers, token and marker compositions, modify the docker-compose-testcase-[1-5*].yml file as needed.

5. To stop the program:

   docker-compose -f docker-compose-testcase-[1-5*].yml down
   

For more detailed information about the implementation, please refer to the REPORT.md file.

Project Files

• main.go: The main file running the token passing and snapshotting logic.
• peerManager.go: Manages peer connections and their relationships.
• queue.go: Implements the FIFO queues for message handling between processes.
• stateManager.go: Manages process states and snapshots.
• tcpCommunication.go: Handles TCP communication and message transmission between peers.
• hostsfile.txt: List of peer hostnames involved in the distributed system.
• Dockerfile: Defines the Docker image for the program.
• docker-compose-testcase-[1-5*].yml: Defines the multi-container Docker application.