The Long Road To Asynchrony

My keynote from PyCon Belarus 2020.

Transcript

1. LONG ROAD ANDREW GODWIN // @andrewgodwin ASYNCHRONY TO THE
2. Andrew Godwin / @andrewgodwin Hi, I’m Andrew Godwin • Django core developer • Worked on Migrations, Channels & Async • Once a Londoner, now from Denver, USA
3. Andrew Godwin / @andrewgodwin
4. Andrew Godwin / @andrewgodwin "Asynchronous Programming" What is it, really?
5. Andrew Godwin / @andrewgodwin Concurrent Programming The more general term
6. Andrew Godwin / @andrewgodwin Input Process Output Sequential execution
7. Andrew Godwin / @andrewgodwin Input Process Output Process Concurrent execution Archive
8. Andrew Godwin / @andrewgodwin Shared use of a single resource In this case, CPUs
9. Andrew Godwin / @andrewgodwin "Communicating Sequential Processes", C. A. R Hoare
10. Andrew Godwin / @andrewgodwin func main() { messages := make(chan string) go func() { messages <- "ping" }() msg := <-messages fmt.Println(msg) }
11. Andrew Godwin / @andrewgodwin Multiple processes multiprocessing Threads threading Event loops asyncio / twisted
12. Andrew Godwin / @andrewgodwin Multiple processes scale best It's also difficult and costs the most!
13. Andrew Godwin / @andrewgodwin Threads are unpredictable Also, the GIL is our ever-present friend
14. Andrew Godwin / @andrewgodwin Event loops are a good compromise They do require shared memory, though.
15. Andrew Godwin / @andrewgodwin Asynchronous ≈ Event loops Most of the time!
16. Andrew Godwin / @andrewgodwin # Ready when a timer finishes await asyncio.sleep(1) # Ready when network packets return await client.get("http://example.com") # Ready when the coroutine exits await my_function("hello", 64.2)
17. Andrew Godwin / @andrewgodwin Network/timer updates An event loop's flow Select a ready task Run task Add new tasks to queue await
18. Andrew Godwin / @andrewgodwin Coroutines Time →
19. Andrew Godwin / @andrewgodwin How did we get here?
20. Andrew Godwin / @andrewgodwin 1998 threading module, Stackless Python 2002 Twisted 2006 Greenlets (later gevent, eventlet) 2008 multiprocessing module 2012 Tulip, PEP 3156 2014 asyncio module 2005 Coroutine-friendly generators (PEP 342)
21. Andrew Godwin / @andrewgodwin 2017 Django Channels 1.0 2018 Django Channels 2.0 2019 DEP 9 (Async support) 2020 Async views land in Django
22. Andrew Godwin / @andrewgodwin No solution is perfect Everyone chooses different tradeoffs
23. Andrew Godwin / @andrewgodwin Asyncio is based on yield from Because it was prototyped in Python 2
24. Andrew Godwin / @andrewgodwin Can't tell if a function returns a coroutine! There are standard hints, but no actual guaranteed way
25. Andrew Godwin / @andrewgodwin async def calculate(x): result = await coroutine(x) return result # These both return a coroutine def calculate(x): result = coroutine(x) return result
26. Andrew Godwin / @andrewgodwin Can't have one function service both How we got here makes sense, but it's still annoying sometimes.
27. Andrew Godwin / @andrewgodwin # Calls get.__call__ instance = MyModel.objects.get(id=3) # Calls get.__call__ # and then awaits its result instance = await MyModel.objects.get(id=3)
28. Andrew Godwin / @andrewgodwin You have to namespace async functions I really, really wish we didn't have to
29. Andrew Godwin / @andrewgodwin instance = MyModel.objects.get(id=3) instance = await MyModel.objects.async.get(id=3)
30. Andrew Godwin / @andrewgodwin Completely different libraries! Even sleep() is different.
31. Andrew Godwin / @andrewgodwin time.sleep ➞ asyncio.sleep requests ➞ httpx psycopg2 ➞ aiopg WSGI ➞ ASGI Django ➞ Django?
32. Andrew Godwin / @andrewgodwin Django & Async
33. Andrew Godwin / @andrewgodwin Asyncio only benefits IO-bound code Code that thrashes the CPU doesn't benefit at all
34. Andrew Godwin / @andrewgodwin We're adding async to some parts The bits where it makes sense!
35. Andrew Godwin / @andrewgodwin But, you can't mix sync and async So we have to have two parallel request paths
36. Andrew Godwin / @andrewgodwin WSGIHandler __call__ WSGI Server WSGIRequest BaseHandler get_response URLs Middleware View __call__ HTTP protocol Socket handling Transfer encodings Headers-to-META Upload file wrapping GET/POST parsing Exception catching Atomic view wrapper Django 3.0 Request Flow
37. Andrew Godwin / @andrewgodwin WSGIHandler __call__ WSGI Server WSGIRequest BaseHandler get_response URLs Middleware Async View __call__ ASGIHandler __call__ ASGI Server ASGIRequest Sync View __call__ Asynchronous request path Proposed async request flow
38. Andrew Godwin / @andrewgodwin WSGIHandler __call__ WSGI Server WSGIRequest URLs Middleware View __call__ ASGIHandler __call__ ASGI Server ASGIRequest Asynchronous request path BaseHandler get_response_async BaseHandler get_response URLs Middleware Async View __call__ Implemented async request flow
39. Andrew Godwin / @andrewgodwin We have to work with what we have I'd rather let people ship code than argue about perfection.
40. Andrew Godwin / @andrewgodwin Django's main job is safety It matters more than anything else
41. Andrew Godwin / @andrewgodwin Deadlocks Livelocks Starvation Race conditions
42. Andrew Godwin / @andrewgodwin Coroutines & the GIL actually help! You're saved from all the awful memory corruption bugs
43. Andrew Godwin / @andrewgodwin async def transfer_money(p1, p2): await get_lock() # Code between awaits is atomic! subtract_money(p1) add_money(p2) await release_lock()
44. Andrew Godwin / @andrewgodwin You can still screw up a lot Trust me, I have lived it while developing async
45. Andrew Godwin / @andrewgodwin async def log_message(m): await client.post("log-server", m) result = calculate_result() log_message(m)
46. Andrew Godwin / @andrewgodwin async def log_message(m): await client.post("log-server", m) result = calculate_result() await log_message(m)
47. Andrew Godwin / @andrewgodwin Async usability has a long way to go But it is undoubtedly the future!
48. Andrew Godwin / @andrewgodwin Python is the language of pragmatism If anyone can get it right, we can
49. Andrew Godwin / @andrewgodwin What does the future hold? Hopefully, no GIL!
50. Andrew Godwin / @andrewgodwin Let's make async understandable Almost every project could benefit, if we made it worth their time.
51. Thanks. Andrew Godwin @andrewgodwin // aeracode.org