distlockd

distlockd is a lightweight, zero-dependency distributed lock server implemented in Python. It enables multiple clients to coordinate access to shared resources using named locks over a simple TCP protocol. Built for simplicity, stability, and ease of integration, distlockd is ideal for teams who need distributed locking without the operational overhead of Redis or other heavyweight systems.

Requirements

• Python >= 3.6

Installation

pip install distlockd

Quick Start

# Start the server
distlockd server --host localhost --port 9001

# In your Python code
from distlockd.client import LockClient

client = LockClient('localhost', 9001)
with client.lock('my-resource'):
    # Your critical section here
    pass

Architecture & Protocol

• Server: Asyncio-based TCP server, in-memory lock management, auto-timeout for stale locks (default: 10s), minimal resource usage.
• Client: Synchronous Python client, context manager support, unique client IDs, retry and timeout mechanisms.
• Protocol: Fast custom binary protocol inspired by RESP (used in Redis). This protocol minimizes parsing overhead and enables sub-millisecond lock operations, making it ideal for high-throughput, low-latency distributed coordination.
• Design Philosophy: Minimalism, reliability, and maintainability. No persistence, no external databases, no complex configuration.

Features

• Named distributed locks (multiple resources, independent lock names)
• Simple TCP-based binary protocol
• Auto-timeout of stale locks (prevents deadlocks)
• Asyncio-based scalable server
• Synchronous, easy-to-use Python client
• Context-manager support for safe lock handling
• Health check endpoint for monitoring
• Configurable timeouts and retry logic
• No external dependencies (no Redis, no databases)
• Structured error handling and logging
• Lightweight and easy to deploy
• Connection pool for efficient connection management

Installation

pip install distlockd

Targeted Use Cases

• Distributed cron jobs: Ensure only one worker runs a scheduled task at a time across multiple hosts.
• Leader election: Elect a leader in a cluster for coordination tasks.
• Resource coordination: Prevent race conditions when accessing shared files, APIs, or other resources.
• Testing and CI: Synchronize test runners or deployment steps in distributed pipelines.
• Lightweight alternatives to Redis locks: When you want distributed locking but find Redis too heavy or complex.
• Temporary critical sections: Where lock persistence is not required and simplicity is key.

Usage

Starting the Server

# Basic usage (default host: 127.0.0.1, port: 9999)
distlockd server

# With custom host and port
distlockd server --host 127.0.0.1 --port 9999

# Enable verbose logging (Note: may impact performance in high-throughput scenarios)
distlockd server -v

Running the Benchmark

distlockd test distlockd|redis

# With custom host and port
distlockd test distlockd|redis --host 127.0.0.1 --port 9999

# With custom iterations, num_clients, num_locks, and throughput_seconds
distlockd test distlockd|redis --iterations 1000 --num_clients 1000 --num_locks 100 --throughput_seconds 10

Client Examples

Basic Usage

from distlockd.client import Client

# Create a Basic client
client = Client() # default host: 127.0.0.1, port: 9999, specified host and port: Client(host="192.xx.xx.xx", port=9999)

# Create a Client with custom timeout and retry logic
client = Client(timeout=5.0, retry=3)

# Create a Client with verbose logging
client = Client(verbose=True)

# Check server health
if client.check_server_health():
    print("Server is healthy!")

Manual Lock Management

# Manual lock acquisition and release
if client.acquire("my-resource", timeout=5.0):
    try:
        print("Lock acquired successfully")
        # Do critical work here...
    finally:
        if client.release("my-resource"):
            print("Lock released successfully")

Using Context Manager

# Using context manager for automatic lock release
try:
    with client.lock("shared-resource", timeout=3.0):
        print("Lock acquired via context manager")
        # Do critical work here...
    # Lock is automatically released when the block exits
    print("Lock automatically released")
except Exception as e:
    print(f"Failed to acquire lock: {e}")

Error Handling

Handle common exceptions:

from distlockd.exceptions import LockAcquisitionTimeout, LockReleaseError, ConnectionError

try:
    with client.lock("resource"):
        # Critical section
        pass
except LockAcquisitionTimeout:
    print("Failed to acquire lock: timeout")
except LockReleaseError as e:
    print(f"Error releasing lock: {e}")
except ConnectionError as e:
    print(f"Connection error: {e}")

Benchmark Comparison Report

Date: 2025-05-28 11:10:51

System Info:

• Platform: Linux
• Platform-Release: 4.18.0-553.5.1.el8_10.x86_64
• Platform-Version: #1 SMP Tue May 21 03:13:04 EDT 2024
• Machine: x86_64
• Processor: x86_64
• CPU Count: 8
• RAM (GB): 15.39
• Uname:
  • system: Linux
  • release: 4.18.0-553.5.1.el8_10.x86_64
  • version: #1 SMP Tue May 21 03:13:04 EDT 2024
  • machine: x86_64
  • processor: x86_64

Test Configuration:

• iterations: 1000 # for latency test
• num_clients: 100 # for concurrency test
• num_locks: 10 # for concurrency test
• throughput_seconds: 10 # for throughput test

Methodology

• Both servers were tested using equivalent client logic and lock contention scenarios.
• Each test reports average, median, and worst-case latency, as well as throughput.
• Tests were run on the same hardware/network for fairness.

🥇 Metric-by-Metric Comparison

✅ Distlockd leads in all latency and throughput metrics 🤝 Both systems maintained a 100% success rate

🧠 Summary

• Distlockd outperforms Redis in every latency and throughput metric, with impressive 28.5k ops/sec.
• Redis remains a solid baseline but is outpaced by Distlockd in high-concurrency locking.
• This makes Distlockd a top candidate for high-performance distributed coordination scenarios.

Acknowledgements

Thanks to Windsurf for providing the benchmarking automation and comparison summary in this project.

License

MIT