Recently, I had the opportunity of deploying django apps that would receive heavy traffic and yet be lite on the memory because I was using a 256 MB virtual private server.
Django has two options:
I have had bad experiences with Mod_Python with issues ranging from memory leaks to difficult configuration. Off late there has been lots of talk about Mod_Wsgi and it's performance.
I decided to do a benchmark test to figure out the truth. For the benchmark test I used Apache 2.2.8 with prefork MPM. I used apache's inbuilt benchamark analyzer ab for the test. The single media file that was being served was a single SWF file with a size of 350KB. The multiple media files being images from a gallery app. Here are the results of a 1000 non-concurrent requests being served:
Total transferred: 4515000 bytes
HTML transferred: 4335000 bytes
Requests per second: 140.90 [#/sec] (mean)
Time per request: 7.097 [ms] (mean)
Time per request: 7.097 [ms] (mean, across all concurrent requests)
Transfer rate: 621.21 [Kbytes/sec] received
Total transferred: 6789000 bytes
HTML transferred: 6595000 bytes
Requests per second: 10.95 [#/sec] (mean)
Time per request: 91.294 [ms] (mean)
Time per request: 91.294 [ms] (mean, across all concurrent requests)
Transfer rate: 72.61 [Kbytes/sec] received
Total transferred: 4513000 bytes
HTML transferred: 4335000 bytes
Requests per second: 158.71 [#/sec] (mean)
Time per request: 6.301 [ms] (mean)
Time per request: 6.301 [ms] (mean, across all concurrent requests)
Transfer rate: 699.43 [Kbytes/sec] received
Total transferred: 6787000 bytes
HTML transferred: 6595000 bytes
Requests per second: 11.05 [#/sec] (mean)
Time per request: 90.487 [ms] (mean)
Time per request: 90.487 [ms] (mean, across all concurrent requests)
Transfer rate: 73.24 [Kbytes/sec] received
Total transferred: 4515000 bytes
HTML transferred: 4335000 bytes
Requests per second: 141.61 [#/sec] (mean)
Time per request: 7.061 [ms] (mean)
Time per request: 7.061 [ms] (mean, across all concurrent requests)
Transfer rate: 624.37 [Kbytes/sec] received
Total transferred: 6789000 bytes
HTML transferred: 6595000 bytes
Requests per second: 10.93 [#/sec] (mean)
Time per request: 91.470 [ms] (mean)
Time per request: 91.470 [ms] (mean, across all concurrent requests)
Transfer rate: 72.47 [Kbytes/sec] received
Total transferred: 4513000 bytes
HTML transferred: 4335000 bytes
Requests per second: 159.35 [#/sec] (mean)
Time per request: 6.275 [ms] (mean)
Time per request: 6.275 [ms] (mean, across all concurrent requests)
Transfer rate: 702.27 [Kbytes/sec] received
Total transferred: 6787000 bytes
HTML transferred: 6595000 bytes
Requests per second: 11.12 [#/sec] (mean)
Time per request: 89.922 [ms] (mean)
Time per request: 89.922 [ms] (mean, across all concurrent requests)
Transfer rate: 73.70 [Kbytes/sec] received
Total transferred: 4472000 bytes
HTML transferred: 4335000 bytes
Requests per second: 344.56 [#/sec] (mean)
Time per request: 2.902 [ms] (mean)
Time per request: 2.902 [ms] (mean, across all concurrent requests)
Transfer rate: 1504.68 [Kbytes/sec] received
Total transferred: 6746000 bytes
HTML transferred: 6595000 bytes
Requests per second: 28.77 [#/sec] (mean)
Time per request: 34.760 [ms] (mean)
Time per request: 34.760 [ms] (mean, across all concurrent requests)
Transfer rate: 189.50 [Kbytes/sec] received
Mod-WSGI gives a 12.5% improvement in performance while serving a single media file while it gives a mere 1% improvement while serving multiple media files. Not much change is seen while apache is made to serve the media, the figures show a small performance improvement. When Nginx comes into the picture it blows away apache to bits. It gives a 116% improvement while serving a single media file and 160% improvement while serving multiple media files. Use Nginx to serve static media (and for reverse proxying) and notice the performance difference.