Q. I can measure network throughput and packet loss using standard UNIX / Linux command line utilities. How do I find out the lateceny and throughput of a web server like Apache under Linux?
A. You need to use the program called httping. It allows you to measure the latency of a webserver and the throughput.
Task: Ping the webserver on host satohost.com
Use the following command for measuring the latency. Press CTRL+c to exit the program. It will display a summary of what was measured.
$ httping -g http://satohost.com
Output:
PING satohost.com:80 (http://satohost.com): connected to satohost.com:80, seq=0 time=981.08 ms connected to satohost.com:80, seq=1 time=709.92 ms connected to satohost.com:80, seq=2 time=1072.02 ms connected to satohost.com:80, seq=3 time=903.81 ms connected to satohost.com:80, seq=4 time=607.84 ms connected to satohost.com:80, seq=5 time=660.01 ms connected to satohost.com:80, seq=6 time=730.12 ms connected to satohost.com:80, seq=7 time=781.49 ms
The -g url option use selects the url to probe / ping. You can also specify the port with -p port option:
$ httping -g http://satohost.com -p 81
You can also connect using SSL, for this to work you need to give a https url or a 443 portnumber:
$ httping -l -g https://satohost.com
OR
$ httping -g http://satohost.com -p 443
Task: Measure throughput of a webserver
The -G option force GET request instead of a HEAD request – this means that also the complete page/file must be transferred. You also need to pass the -b option with -G option to get the transferspeed (in KB/s).
$ httping -Gbg http://satohost.com/
Output:
PING satohost.com:80 (http://satohost.com/): connected to satohost.com:80, seq=0 time=1738.39 ms 22KB/s connected to satohost.com:80, seq=1 time=1650.19 ms 20KB/s connected to satohost.com:80, seq=2 time=1759.65 ms 17KB/s connected to satohost.com:80, seq=3 time=1589.98 ms 21KB/s connected to satohost.com:80, seq=4 time=3709.87 ms 6KB/s connected to satohost.com:80, seq=5 time=3329.69 ms 7KB/s .... ..... --- http://satohost.com/ ping statistics --- 53 connects, 53 ok, 0.00% failed round-trip min/avg/max = 1451.9/2013.6/11656.0 ms Transfer speed: min/avg/max = 6/19/24 KB
Please note above in above command you’re no longer measuring the latency!
You can also pass -X option with -G to show the amount of data transferred (excluding the headers):
$ httping -XGbg http://satohost.com/
Output:
PING satohost.com:80 (http://satohost.com/): connected to satohost.com:80, seq=0 time=1576.11 ms 22KB/s 19KB connected to satohost.com:80, seq=1 time=2620.26 ms 9KB/s 19KB connected to satohost.com:80, seq=2 time=1507.69 ms 23KB/s 19KB connected to satohost.com:80, seq=3 time=1522.08 ms 24KB/s 19KB connected to satohost.com:80, seq=4 time=1533.68 ms 23KB/s 19KB connected to satohost.com:80, seq=5 time=1581.92 ms 21KB/s 19KB connected to satohost.com:80, seq=6 time=1512.06 ms 24KB/s 19KB --- http://satohost.com/ ping statistics --- 7 connects, 7 ok, 0.00% failed round-trip min/avg/max = 1507.7/1693.4/2620.3 ms Transfer speed: min/avg/max = 9/21/24 KB
Test remote server CPU
The -B option along with -G option ask the HTTP server to compress the returned data – this will reduce the influence of the bandwidth of your connection while increasing the influence of the processorpower of the HTTP server.
$ httping -BGg http://satohost.com/
Flood Webserver
The -f option used to flood ping i.e. do not sit idle between each ping but ping as fast as the computer and network allow you to (don’t run this on production or 3rd party servers):
$ httping -fg http://satohost.com/
(adsbygoogle = window.adsbygoogle || []).push({});
