Architecture / 01STARTER

ASGI vs WSGI & The Uvicorn Event Loop

“WSGI (Web Server Gateway Interface) is synchronous: each concurrent HTTP connection occupies an entire OS worker thread. ASGI (Asynchronous Server Gateway Interface) is asynchronous: a single Python process running an event loop (Uvicorn on uvloop) coordinates thousands of concurrent requests by suspending I/O waiting periods via standard async/await.”

Domain Category

ASGI & Starlette Core

Why traditional WSGI synchronous servers block threads, and how ASGI enables non-blocking async Python I/O.

Physical World Analogy

A restaurant waiter (WSGI) who stands by the table for 15 minutes waiting while the chef cooks the meal, serving only 4 tables an hour vs an ASGI waiter who takes 50 orders in 2 minutes, passing tickets to the kitchen and delivering dishes as soon as they are ready.

FastAPI Implementation Contract

Python • Pydantic V2 • Async Endpoint Code

main.py (FastAPI 0.110+)RUST PYDANTIC V2

# ASGI Application Function Contract
async def app(scope, receive, send):
    assert scope['type'] == 'http'
    await send({
        'type': 'http.response.start',
        'status': 200,
        'headers': [[b'content-type', b'application/json']],
    })
    await send({
        'type': 'http.response.body',
        'body': b'{"status": "ok", "engine": "uvicorn"}',
    })

Low-level ASGI protocol interface processing raw HTTP sockets.

Execution Lifecycle

ASGI & Event Loop Pipeline Flow

01STEP

Uvicorn opens non-blocking socket listener via uvloop (libuv C binding)

02STEP

Constructs ASGI scope dictionary containing request headers, client IP, and HTTP path

03STEP

Calls application(scope, receive, send) asynchronously

04STEP

Middleware intercepts and mutates scope/headers

05STEP

Streams response bytes chunk-by-chunk through send({"type": "http.response.body"})

Concurrency Traps & Loop Starvation

Calling blocking synchronous libraries (e.g. time.sleep() or requests.get()) inside an async def endpoint (starves the entire event loop)

Running Uvicorn with a single worker in production without a Gunicorn process manager

Systems Engineering Optimization

Running Uvicorn with uvloop delivers 300% higher HTTP request throughput than standard Python asyncio by utilizing libuv native C event polling.

Related Backend Concepts

Concurrency / 01

ASGI vs WSGI & The Uvicorn Event Loop

A restaurant waiter (WSGI) who stands by the table for 15 minutes waiting while the chef cooks the meal, serving only 4 tables an hour vs an ASGI waiter who takes 50 orders in 2 minutes, passing tickets to the kitchen and delivering dishes as soon as they are ready.

Python • Pydantic V2 • Async Endpoint Code

ASGI & Event Loop Pipeline Flow

Related Backend Concepts

Endpoint Execution: async def vs def (ThreadPool Offload)

Pydantic V2 & Rust Core Serialization Engine

Hierarchical Dependency Injection (DAG & Yield Contexts)

ASGI vs WSGI & The Uvicorn Event Loop

A restaurant waiter (WSGI) who stands by the table for 15 minutes waiting while the chef cooks the meal, serving only 4 tables an hour vs an ASGI waiter who takes 50 orders in 2 minutes, passing tickets to the kitchen and delivering dishes as soon as they are ready.

Python • Pydantic V2 • Async Endpoint Code

ASGI & Event Loop Pipeline Flow

Related Backend Concepts

Endpoint Execution: async def vs def (ThreadPool Offload)

Pydantic V2 & Rust Core Serialization Engine

Hierarchical Dependency Injection (DAG & Yield Contexts)