How to handle sockets - easily?

William Ahern william at wilbur.25thandClement.com
Wed Feb 16 19:59:27 CET 2011


Bubba <nickname at banelli.biz.invalid> wrote:
<snip>
> import asyncore
> import socket
> import string
> import MySQLdb
> import sys
<snip>
>     def __init__(self, host, port):
>         asyncore.dispatcher.__init__(self)
>         self.create_socket(socket.AF_INET, socket.SOCK_STREAM)
>         self.set_reuse_addr()
>         self.bind((host, port))
>         self.listen(5)
>
>     def handle_accept(self):
>         pair = self.accept()
>         if pair is None:
>             pass
>         else:
>             sock, addr = pair
>             print 'Incoming connection from %s' % repr(addr)
>             handler = Handler(sock)
>
> server = Server('', 2020)
> asyncore.loop()
>
<snip>
> I do, however, have some more questions (thus crosspost to
> comp.lang.python) - how many connections can this server handle without
> a problem? I'm using Asyncore module, as it can be seen. 

> Is it necessary, due to the fact that it should serve more than
> thousand devices that send data every 10 seconds, to do threading (I
> believe that is already done with Asyncore for sockets, but what about
> SQL?) 

The MySQL C library is not asynchronous. Each request does blocking I/O. If
the MySQLdb module is a wrapper for the MySQL C library--and it seems it
is--then you will want to use threads (not coroutines or generators). For
all I know your accept handler is threaded already. MySQL itself almost
certainly can't handle tens of thousands of simultaneous requests. The
backend connection and query handler is simply-threaded as well, which means
for every connection your talking 2 * "tens of thousands" threads. I didn't
read over your code much, but the only way to get around this would be to
handle your socket I/O asynchronously; but I don't enough about Python to
get the mixed behavior you'd want.

I think that there's an asynchronous all-Python MySQL library, but I'm not
sure. Maybe one day I can open source my asynchronous MySQL C library. (I
always recommend people to use PostgreSQL, though; which is superior in
almost every way, especially the C client library and the wire protocol.)



More information about the Python-list mailing list