Memberly

Contributors

Pruthvi Prakash Navada

Overview

This project implements a distributed peer-to-peer system with membership management, leader election, and fault tolerance. The system utilizes both TCP and UDP for communication, implements heartbeat-based failure detection, and maintains consistency through a view-based approach.

Directory Structure

.
├── Dockerfile
├── docker-compose-testcase-1.yml
├── docker-compose-testcase-2.yml
├── docker-compose-testcase-3.yml
├── docker-compose-testcase-4.yml
├── hostsfile.txt
├── main.go
└── peer/
    ├── config/
    │   └── config.go
    ├── datastructures/
    │   ├── safelist.go
    │   └── safevalue.go
    ├── handlers/
    │   └── messagehandler.go
    ├── network/
    │   ├── connection.go
    │   └── peer.go
    ├── types/
    │   └── types.go
    └── utils/
        └── utils.go

Building and Running

Prerequisites

• Docker
• Docker Compose
• Go 1.18 or higher (for local development)

Building the Container

# Build the Docker image
docker build -t prj3 .

Running Test Cases

The system includes four different test scenarios, each with its own docker-compose file:

# Test Case 1: Basic Join Operations
docker-compose -f docker-compose-testcase-1.yml up

# Test Case 2: Single Node Failure
docker-compose -f docker-compose-testcase-2.yml up

# Test Case 3: Multiple Node Failures
docker-compose -f docker-compose-testcase-3.yml up

# Test Case 4: Leader Failure with Message Loss
docker-compose -f docker-compose-testcase-4.yml up

Command Line Arguments

The program accepts the following command-line flags:

• -h string: Path to the hosts file (required)
• -d int: Initial delay in seconds (default: 0)
• -c int: Crash delay in seconds
• -t bool: Trigger crash simulation

System Architecture

Network Communication

• TCP Port: 8080 (membership management)
• UDP Port: 8081 (heartbeat messages)
• Heartbeat Interval: 1.1 seconds
• Monitor Interval: 2.2 seconds

Key Features

1. Membership Management

   • View-based consistency
   • Quorum-based acknowledgments
   • Member addition/removal

2. Failure Detection

   • Heartbeat-based monitoring
   • UDP broadcast for efficiency
   • Configurable monitoring intervals

3. Leader Election

   • Automatic leader election on failure
   • State synchronization after election
   • Pending operation recovery

Implementation Notes

Important Considerations

1. Network Configuration

   • Ensure ports 8080 and 8081 are available
   • Docker network must support UDP broadcast
   • Hostname resolution must be functional

2. Safety Mechanisms

   • Thread-safe data structures
   • Mutex-protected state changes
   • Atomic view updates

3. Failure Handling

   • Graceful handling of network partitions
   • Recovery from message loss
   • State consistency maintenance

Error Conditions

1. Network Errors

   Error starting TCP listener: address already in use
   Error reading from UDP connection: connection refused
   Error getting hostname from address: no such host
   

2. State Errors

   Error: member ID not found in peerIdToName map
   Error encoding integer: invalid argument
   Error decoding byte array to integers: unexpected EOF
   

Testing

The system includes four test scenarios:

1. Test Case 1: Basic membership protocol

   • Tests join operations
   • Verifies view synchronization
   • Validates message handling

2. Test Case 2: Single node failure

   • Tests failure detection
   • Verifies member removal
   • Validates state updates

3. Test Case 3: Multiple node failures

   • Tests multiple node failures
   • Verifies member removal
   • Validates state updates

4. Test Case 4: Leader failure with message loss

   • Tests leader failure handling
   • Verifies message loss recovery
   • Validates consistency maintenance