Performance comparison of Django's own WSGI server vs Deployment Uwsgi+nginx

First, let's take a look at the test cloud host

Cpu:

[Email protected]:/tmp/webbench-1.5cat /proc/cpuinfo | grep Modelmodel          the model name    : Intel (R) Core (TM) i5-4260U1. 40ghz[email protected]:/tmp/ webbench-1.5#

Mem:

[Email protected]:/tmp/webbench-1.5 free-m             total       used       free      shared    buffers     cachedmem:           490                 364          3          a        103-/+ Buffers/cache:        219        Swap:          509         208        301

Two. Test comparison

The test tool is webbench.

• The following is the performance of Django's own Wsgi webserver

[Email protected]:/tmp/webbench-1.5# webbench-c --T -http//115.28.21.25:8008/Webbench-simple Web Benchmark1.5Copyright (c) Radim Kolar1997-2004, GPL Open Source Software.Benchmarking:GET http://115.28.21.25:8008/ -Clients, running -sec. speed=554Pages/min,129959bytes/sec. Requests:277Susceed,0Failed.

will probably be able to withstand the 20-way concurrency, and when it goes up to 25, a failed request begins.

• The following replaces the WSGI server with Uwsgi+nginx

[Email protected]:/tmp/webbench-1.5# webbench-c the-T -http//115.28.21.25:8000/Webbench-simple Web Benchmark1.5Copyright (c) Radim Kolar1997-2004, GPL Open Source Software.Benchmarking:GET http://115.28.21.25:8000/ theClients, running -sec. speed=640Pages/min,126746bytes/sec. Requests: theSusceed,0Failed.

Probably able to withstand the 170-way concurrency, the effect is obvious, when the request continues to rise, will start to appear the failure request.

Performance comparison of Django's own WSGI server vs Deployment Uwsgi+nginx