From: Jay Levitt <jay+news@...>
Date: 2007-10-04T05:20:11+09:00
Subject: Re: Recent Criticism about Ruby (Scalability, etc.)

On Thu, 4 Oct 2007 03:06:03 +0900, Gary Wright wrote:

> I admit to being puzzled by the general fascination
> with threads (native or not) for solving scaling
> problems.
> 
> Good old processes have always seemed like a reasonable
> way to partition problems and take advantage of
> concurrency opportunities due to waiting on I/O
> (single processor) or the parallel nature of a
> CPU-bound calculation (multi-processor).

Ayup.  At AOL, when we grew beyond the abilities of sendmail to handle
incoming mail from the Internet, we rolled our own SMTP server.  It had the
following characteristics:

- Non-threaded, event-driven main loop
- Asynchronous DNS calls
- Asynchronous database calls
- No disk I/O other than logging (which itself might have been async, I
forget)
- Therefore, no process ever waited on anything

The best configuration?  One process per CPU.


-- 
Jay Levitt                |
Boston, MA                | My character doesn't like it when they
Faster: jay at jay dot fm | cry or shout or hit.
http://www.jay.fm         | - Kristoffer