Summary

Python has had support for asynchronous programming since version 1.5.2, with the introduction of the asyncore and asynchat modules for asynchronous network programming. Version 2.5 introduced the ability to send data to coroutines via yield expressions, allowing us to write asynchronous code in a simpler but more powerful way. Python 3.4 introduced a new library for asynchronous I/O called asyncio.

Python 3.5 introduced true coroutine types via async def and await. Interested readers are encouraged to explore these new developments. One word of warning though: asynchronous programming is a powerful tool that can dramatically improve the performance of I/O-intensive code. It does not come without issues, though, the main of which is complexity.

Any important asynchronous code has to carefully select nonblocking libraries in order to avoid using blocking code. It has to implement a coroutine scheduler (since the OS does not schedule coroutines for us like it does with threads), which involves writing an event loop and adding further complexity. Reading asynchronous code can be challenging to the point that even our simple examples do not look all that simple at first sight. Handle with care!