[6767] in www-talk@info.cern.ch
More on HTTP-NG timings
daemon@ATHENA.MIT.EDU (Simon E Spero)
Tue Nov 22 11:58:44 1994
Date: Tue, 22 Nov 1994 17:12:58 +0100
Errors-To: listmaster@www0.cern.ch
Reply-To: ses@tipper.oit.unc.edu
From: Simon E Spero <ses@tipper.oit.unc.edu>
To: Multiple recipients of list <www-talk@www0.cern.ch>
Rob - the prelim tests I did were based on fetching lots of small documents
from unc to HP labs in Bristol. The bandwidth was estimated by fetching a large
file (2Mb) using raw TCP.
I've also done some more tests this morning (when the network connection is
less busy). To get a more realistic idea of real-world performance, I used
as a data set the sunsite home page, and one instance of each of its inline
images. I ran the servers on tipper to eliminate any noise from the sunsite
net load (it's a different T1 to the campus one). Total size was ~48K, split
up between 5 files.
HTTP 1.0 gave an average transfer time of 24.7 seconds. HTTP-NG 0.9 had an
average time of around 5 seconds
The same test, using a machine at Ventana Media (In NC, but routed via DC),
gave an average time for HTTP-TOS of around 5.6 seconds, with HTTP-NG
clocking around 1.5 seconds.
Simon
------- End of Forwarded Message
