A case of client that affects the speed of the server
| From: | Carlos Henrique Barbosa Gon�alves | Date: | Fri, 19 Oct 2001 19:47:51 +0000 |
| Subject: | A case of client that affects the speed of the server | ||
| Groups: | php.general | ||
| Request: | Send a blank email to php-general+get-71620@lists.php.net to get a copy of this message | ||
Hello, people.
I'm trying to configure Apache+PHP on a Linux Red Hat machine(Pentium III, IP 130.94.20.132) in order to leave the machine I'm currently using (Pentium II, www.etapa.net)
To evaluate the difference in performance, I ran the same PHP script in both machines. Much to my surprise, the machine with a better hardware was slower.
It's been some 10 days I've been working to fix this. Perhaps you can help me.
At this moment, both servers are running the same version of Apache and PHP. You can check this through the addresses:
www.etapa.net/info.php3
and
130.94.20.132/info.php3
The script I'm using to test the difference in performance can be reached through
www.etapa.net/enem_pontosnew.php3
and
130.94.20.132/enem_pontosnew.php3
If you want to see the script source, use 130.94.20.132/enem_pontosnew.txt and see the source.
In both of them you will see five measurements of time. My readings, from a lynx browser at the USA are
For www.etapa.net,
0.029524922370911
0.030010938644409
0.031177997589111
0.047765970230103
0.064495921134949
For 130.94.20.132
0.0050849914550781
0.0051850080490112
0.0054060220718384
0.024041056632996
0.038671016693115
If you look closely at the above figures, you'll notice that in the first three readings of each series, the new server is consistently 6 times faster than the old one (which of course is something very good). However, in the last two readings, the new server is only two times faster than the old one. During the assembling of the html code of the tables in the pages, the new server reduces its performance in a significant way.
Now an even more unexpected result. From an IE browser in Brasil, I've been getting approximately the same readings from etapa.net, but 130.94.20.132 is slower than the same page accessed from the USA. See it:
0.014531970024109
0.014856934547424
0.015432953834534
0.5794450044632
1.1874959468842
My conclusion is that the speed with which the browser can exchange bytes with the server at the 130.94.20.132 address is affecting the speed of the server.
How should I configure the server so that it is independent of the client? (Unfortunately, the people that have configured my other server don't know the answer for this question.)
Thanks.
_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp