ICS 4104: Distributed Systems - Load Balancer Assignment

Overview

This project implements a customizable load balancer using consistent hashing to distribute client requests across multiple server containers. The system is built with Python (Flask), Docker, and a Docker network for communication. It supports dynamic scaling, failure recovery, and load distribution analysis.

Design Choices

• Language: Python with Flask for simplicity and robust HTTP handling.
• Consistent Hashing: Implemented using a sorted list with linear probing for conflict resolution.
• Docker: Containers for servers and load balancer, managed via docker-compose.
• Failure Detection: Heartbeat checks every 5 seconds to detect and replace failed servers.
• Hash Functions: Used as specified: H(i) = i + 2i + 2172 for requests, Φ(i,j) = i + j + 2j + 25 for virtual servers.

Assumptions

• Docker is installed and running on Parrot OS (Debian-based).
• Servers are stateless, handling only /home and /heartbeat endpoints.
• Random request IDs are used for hashing to simulate client requests.
• Hostnames are unique and randomly generated if not specified.

Setup and Deployment

1. Install Docker:

   sudo apt-get update
   sudo apt-get install ca-certificates curl gnupg lsb-release
   sudo mkdir -p /etc/apt/keyrings
   curl -fsSL https://download.docker.com/linux/debian/gpg | sudo gpg --dearmor -o /etc/apt/keyrings/docker.gpg
   echo "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/debian $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
   sudo apt-get update
   sudo apt-get install docker-ce docker-ce-cli containerd.io docker-compose-plugin

2. Clone Repository:

   git clone <your-repo-url>
   cd load-balancer

3. Build and Run:

   make build
   make up

4. Access Load Balancer: Endpoints are available at http://localhost:5000.

Testing

Run the test script to verify functionality:

make test

Test Script

The test script (tests/test_load_balancer.py) sends 10,000 async requests, tests scaling, and simulates server failure.

import asyncio import aiohttp import matplotlib.pyplot as plt import random import json

async def send_request(session, url): async with session.get(url) as response: return await response.json()

async def test_load_distribution(): async with aiohttp.ClientSession() as session: counts = {'server_1': 0, 'server_2': 0, 'server_3': 0} tasks = [send_request(session, 'http://localhost:5000/home') for _ in range(10000)] responses = await asyncio.gather(*tasks) for resp in responses: server = resp['message'].split(': ')[1] counts[server] = counts.get(server, 0) + 1

    plt.bar(counts.keys(), counts.values())
    plt.title('Request Distribution Across 3 Servers')
    plt.savefig('load_distribution.png')
    print("A-1: Request counts:", counts)

async def test_scaling(): async with aiohttp.ClientSession() as session: loads = [] for n in range(2, 7): # Add servers payload = {'n': n - 3, 'hostnames': [f'server_{i}' for i in range(4, n+1)]} if n > 3 else {'n': 0} async with session.post('http://localhost:5000/add', json=payload) as resp: pass counts = {} tasks = [send_request(session, 'http://localhost:5000/home') for _ in range(10000)] responses = await asyncio.gather(*tasks) for resp in responses: server = resp['message'].split(': ')[1] counts[server] = counts.get(server, 0) + 1 avg_load = sum(counts.values()) / len(counts) loads.append(avg_load) # Reset to 3 servers if n > 3: async with session.delete('http://localhost:5000/rm', json={'n': n-3, 'hostnames': [f'server_{i}' for i in range(4, n+1)]}) as resp: pass plt.plot(range(2, 7), loads) plt.title('Average Load vs Number of Servers') plt.savefig('scaling.png') print("A-2: Average loads:", loads)

async def test_failure_recovery(): async with aiohttp.ClientSession() as session: # Simulate failure by stopping server_1 import docker client = docker.from_env() client.containers.get('server_1').stop() await asyncio.sleep(10) # Wait for replacement async with session.get('http://localhost:5000/rep') as resp: data = await resp.json() print("A-3: Replicas after failure:", data['message']['replicas'])

if name == 'main': asyncio.run(test_load_distribution()) asyncio.run(test_scaling()) asyncio.run(test_failure_recovery())