CODEWITHSUNDEEP
pythonconcurrencyoperating-systempython-asynciomutex
abraguez
abraguez

Reputation: 375

Python Asyncio preferred concurrent design for scalable application

I'm trying to build an appllication with python's asyncio module that is scalable and uses other modules based on asyncio also, the idea is to easily add tasks as the application grows, utilizing syncronization primitives for shared resources between the tasks, yet I'm confused regarding which design would best fit the intent.

import asyncio

async def task1():
    while True:
        # Code from task1

        await asyncio.sleep(1)

async def task2():
    while True:
        # Code from task2

        await asyncio.sleep(1)

async def task3():
    while True:
        # Code from task3
        
        await asyncio.sleep(1)



async def main():
    await asyncio.gather(
        task1(),
        task2(),
        task3()
    )

if __name__ == "__main__":
    asyncio.get_event_loop().run_until_complete(main())

In the above approach each task is a while loop, yielding at the end of each execution,

import asyncio

async def task1():
    # Code from task1

    await asyncio.sleep(1)

async def task2():
    # Code from task2

    await asyncio.sleep(1)

async def task3():
    # Code from task3
    
    await asyncio.sleep(1)



async def main():
    while True:
        await asyncio.gather(
            task1(),
            task2(),
            task3()
        )

if __name__ == "__main__":
    asyncio.get_event_loop().run_until_complete(main())

In this last approach, the gather method is the one being inside a while loop.

I also that perhaps wouldn't be necessary as asyncio tasks are running in a cooperative way, and I would get the same result from awaiting in series each coroutine, saving the usage of mutexes.

Upvotes: 0

Views: 136

Answers (1)

Frank Yellin
Frank Yellin

Reputation: 11307

Is there anything asynchronous happening besides the asyncio.sleep(1). If not, then why bother using asynchronicity.

In the first piece of code, your three tasks are running independently of each other; you might handle 3 items in task1, 4 items in task2, and none in task3.

In the second, you specifically handle one thing in task1, one thing in task2, one thing in task3, and then start over again. If the tasks run at different speeds, this seems inefficient.

Upvotes: 0

Related Questions