Async file write with tornado

app.py

import os
from uuid import uuid4

from tornado import web
from tornado.httpserver import HTTPServer
from tornado.ioloop import IOLoop
from tornado.iostream import PipeIOStream

IOLoop.configure('tornado.platform.asyncio.AsyncIOMainLoop')


class SafeFileIOStream:
    def __init__(self, fname):
        self.fname = fname

    def __enter__(self):
        # Create file
        os.open(self.fname, os.O_CREAT)
        # Create stream
        fd = os.open(self.fname, os.O_WRONLY)
        self.stream = PipeIOStream(fd)
        return self.stream

    def __exit__(self, exc_type, exc_val, exc_tb):
        # Close stream
        self.stream.close()


class TestHandler(web.RequestHandler):
    async def get(self):
        fname = '/tmp/' + uuid4().hex
        with SafeFileIOStream(fname) as stream:
            await stream.write(b'hello stranger')
        self.write('success')


app_handlers = [
    (r'/test/?', TestHandler),
]

application = web.Application(app_handlers)

# application.listen(8000, '0.0.0.0')

if __name__ == '__main__':
    server = HTTPServer(application)
    server.bind(8000)
    server.start(0)

    loop = IOLoop.current()
    loop.start()

ab -c 8 -n 8000 http://127.0.0.1:8000/test/

This is ApacheBench, Version 2.3 <$Revision: 1748469 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking 127.0.0.1 (be patient)
Completed 800 requests
Completed 1600 requests
Completed 2400 requests
Completed 3200 requests
Completed 4000 requests
Completed 4800 requests
Completed 5600 requests
Completed 6400 requests
Completed 7200 requests
Completed 8000 requests
Finished 8000 requests

Server Software: TornadoServer/4.4.1
Server Hostname: 127.0.0.1
Server Port: 8000

Document Path: /test/
Document Length: 7 bytes

Concurrency Level: 8
Time taken for tests: 9.722 seconds
Complete requests: 8000
Failed requests: 0
Total transferred: 1608000 bytes
HTML transferred: 56000 bytes
Requests per second: 822.86 #/sec
Time per request: 9.722 ms
Time per request: 1.215 [ms](mean, across all concurrent requests)
Transfer rate: 161.52 [Kbytes/sec] received

Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 0.2 0 8
Processing: 2 9 10.3 8 548
Waiting: 2 9 10.2 8 548
Total: 2 10 10.4 8 550

Percentage of the requests served within a certain time (ms)
50% 8
66% 9
75% 10
80% 11
90% 14
95% 19
98% 28
99% 36
100% 550 (longest request)

This may seem async, because the server is using multiple processes, thus it executes requests in parallel. If you limit the processes to one (server.start(1)), the server will be sync and not asyc. Also, try using bigger files.