Atlas

Atlas is a reverse proxy based load balancing framework implemented in Python. It exposes various load balancing algorithms to distribute incoming requests across multiple origin servers. The load balancer state is maintained in-memory at the reverse proxy server, with no (or minimal) set up needed on the origin server.

Installation

Atlas can be installed from the Python Package Index using the following command:

pip install atlaslb

Usage

Supported Algorithms

1. Random
2. Round Robin
3. Least Connection
4. Least Latency

OriginServer

The OriginServer class represents an origin server with various attributes such as host, weight, CPU usage, local requests in flight (local_rif), and latency.

Methods

• __init__(self, host, weight=1, cpu=0, local_rif=0, latency=0): Initializes the origin server with the given attributes.
• __str__(self): Returns a string representation of the origin server.

BaseLoadBalancer

The BaseLoadBalancer class is the base class for all load balancers. It maintains a list of origin servers and a mapping of hostnames to server objects.

Methods

• __init__(self, servers: dict[str, OriginServer] = None): Initializes the load balancer with the given servers.
• add_server(self, server: OriginServer): Adds a server to the load balancer.
• forward_request(self, request, path: str): Forwards the request to the appropriate origin server based on the load balancing algorithm.
• __str__(self): Returns a string representation of the load balancer.

LoadBalancerFactory

The LoadBalancerFactory class is a factory class to create different types of load balancers based on the given configuration.

Methods

• CreateLoadBalancer(self, configuration): Creates a load balancer based on the provided configuration.