I guess what I am trying to figure out is how much is pulled from the httpd.conf for optimizations if any, how much is over-ruled by the LiteSpeed web interface if any and more detailed information on how they interact.
I set the MaxClients and ListenBacklog in httpd.conf both to 1 and then got the following results with LiteSpeed vs Apache:
----------------------------
The first results are with LiteSpeed and the httpd.conf of apache set to MaxClients 1 ListenBacklog 1
The second results are with Apache and the httpd.conf of apache set to MaxClients 1 ListenBacklog 1
michael@mikes-terminal ~/Documents/litespeed_license $ ab -c 20 -n 1000
http://127.0.0.1/~test/
This is ApacheBench, Version 2.3 <$Revision: 655654 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd,
http://www.zeustech.net/
Licensed to The Apache Software Foundation,
http://www.apache.org/
Benchmarking 127.0.0.1 (be patient)
Completed 100 requests
Completed 200 requests
Completed 300 requests
Completed 400 requests
Completed 500 requests
Completed 600 requests
Completed 700 requests
Completed 800 requests
Completed 900 requests
Completed 1000 requests
Finished 1000 requests
Server Software: LiteSpeed
Server Hostname: 127.0.0.1
Server Port: 80
Document Path: /~test/
Document Length: 839 bytes
Concurrency Level: 20
Time taken for tests: 1.151 seconds
Complete requests: 1000
Failed requests: 0
Write errors: 0
Total transferred: 1004000 bytes
HTML transferred: 839000 bytes
Requests per second: 868.88 [#/sec] (mean)
Time per request: 23.018 [ms] (mean)
Time per request: 1.151 [ms] (mean, across all concurrent requests)
Transfer rate: 851.91 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 4 10 1.9 10 15
Processing: 7 13 2.2 13 26
Waiting: 7 13 2.2 13 26
Total: 13 23 1.7 23 32
Percentage of the requests served within a certain time (ms)
50% 23
66% 23
75% 23
80% 24
90% 25
95% 26
98% 27
99% 28
100% 32 (longest request)
-----------------------------------------------
michael@mikes-terminal ~/Documents/litespeed_license $ ab -c 20 -n 1000
http://127.0.0.1/~test/This is ApacheBench, Version 2.3 <$Revision: 655654 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd,
http://www.zeustech.net/
Licensed to The Apache Software Foundation,
http://www.apache.org/
Benchmarking 127.0.0.1 (be patient)
Completed 100 requests
Completed 200 requests
apr_socket_recv: Connection reset by peer (104)
Total of 275 requests completed
--------------
I then set the Max Connections to 1 in the litespeed configuration and repeated the litespeed test which resulted in approximately the same results as the first test.
I then set all the keepalive, caching, and buffering options to 0 that I could find and still it completed all 1000 requests with no failures.
This is not really clear either way as to what is going on, I theorize that the LiteSpeed equivalent of the ListenBacklog setting of Apache is allowing the connections to queue up and be completed, which if that theory is correct would mean that LiteSpeed is not using the httpd.conf for that optimization.
So what I am really trying to find out is how LiteSpeed's configuration and the httpd.conf interact.