Glossary: concurrency, conditions and containers — DedSec Project

Concurrency

Concept in practice

Concurrency means multiple tasks can make progress during overlapping periods; they need not execute simultaneously. Parallelism specifically means execution at the same time on multiple processing resources. Interleaved tasks can still share mutable state and create ordering hazards.

Focused example

import asyncio
async def job(i):
    await asyncio.sleep(0)
    return i * 2
async def main():
    return await asyncio.gather(job(1), job(2))
assert asyncio.run(main()) == [2, 4]

Exercise and common pitfall

Verify that gather returns results in the same input order even if individual jobs complete at different times. Contrast cooperative I/O waits with CPU-heavy work that blocks the event loop; do not assume async code automatically runs on multiple CPU cores.

Reference: https://docs.python.org/3/library/asyncio-task.html

Condition

Concept in practice

A conditional chooses a branch using a predicate. Program correctness depends not just on truthiness but on the exact boundary conditions: greater-than differs from greater-than-or-equal, and None, empty strings or zero may be coerced to false in ways that obscure meaning.

Focused example

def shipping_cost(total):
    if total < 0: raise ValueError('negative total')
    if total >= 50: return 0
    return 5
assert shipping_cost(49.99) == 5
assert shipping_cost(50) == 0

Exercise and common pitfall

Test exactly below, at and above the 50-unit threshold. Change the order of conditions and explain whether a negative input could be treated as a normal shipping price. Prefer explicit is None when missing data differs from numeric zero.

Reference: https://docs.python.org/3/tutorial/controlflow.html#if-statements

Container

Concept in practice

An application container packages processes with filesystem layers and runtime settings while relying on the host kernel for isolation. A container is not a virtual machine: it normally shares the host kernel, and its network, users, capabilities and mounts must be configured intentionally.

Focused example

# On a machine with Docker and a trusted base image:
docker run --rm --network none --read-only alpine:3.20 sh -c 'id; uname -s'

Exercise and common pitfall

On a supported host, compare uname inside and outside the container, then inspect which file writes fail with --read-only. Never mount your home directory or container engine socket just to make a demo succeed; such mounts change the security boundary.

Reference: https://docs.docker.com/engine/security/

Coroutine

Concept in practice

A coroutine is a computation whose execution can suspend and later resume. In Python, calling an async def function creates a coroutine object; it is not automatically executed to completion. An event loop orchestrates awaits and scheduled tasks.

Focused example

import asyncio
async def answer():
    await asyncio.sleep(0)
    return 42
assert asyncio.run(answer()) == 42

Exercise and common pitfall

Try created = answer() without awaiting and inspect its type; then explicitly close it if the experiment does not await it, so no warning is leaked. Inside an already running event loop use await, not a nested asyncio.run.

Reference: https://docs.python.org/3/library/asyncio-task.html