On Thu, Feb 26, 2015 at 12:00 AM, Antoine Pitrou <[email protected]> wrote:
> On Wed, 25 Feb 2015 23:44:33 +0100 > Ludovic Gasc <[email protected]> wrote: > > > > I've disabled keep_alive in api_hour, I quickly tested on agents list > > webservices via localhost, I've 3334.52 req/s instead of 4179 req/s, > 0.233 > > latency average instead of 0.098 and 884 errors instead of 0 errors. > > It isn't a big change compare to others Web frameworks values, but it's a > > change. > > IMO, the fact that you get so many errors indicates that something is > probably wrong in your benchmark setup. If you have a technical clue, I'm interested in. > It is difficult to believe that > Flask and Django would believe so badly in such a simple (almost > simplistic) workload. > Sorry guys, but often, things mustn't be complicated to be good: I've filled the requirements of my client with this WebService. As a Computer Science engineer, I've learnt at school that to prove my skills, I need to build complicated solutions. For me, it's like to shot in your foot: Yes, you prove your value, but finally, you need to change the business logic several times, you need to maintain code. If it's very complicated at the beginning, few chances it will be simpler at the end. For example, see the Web frameworks: more and more layers to try to assist developers, but, finally, not really sure that everybody saves time in the story. BTW, I want to publish a more complete version of this daemon used in the benchmark but almost nobody can test easily, the complete daemon use Panoramisk to discuss with Asterisk. The setup is a little bit more complicated that a classical Web architecture. FYI, I've in the pipe a benchmark to compare an AGI daemon built with Panoramisk and xivo-agid. > > Regards > > Antoine. > > >
