cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Berin Loritsch <>
Subject [C2] Performance with Cache enabled
Date Fri, 04 May 2001 16:00:44 GMT
My friends, I want to say Kudos to all of the Cache implementors.  I ran my
typical Apache JMeter test run on my machine, and Cocoon was still breathing
easily.  So I decided to go nuts.

My Test System:

Athlon 750 with 256MB RAM
Windows 2000
Tomcat 3.3b2
Cocoon 2 (CVS as of 11:40 AM 5/4/2001)
Sun JDK 1.3.0_01 (with -Xmx256M option)
Sun HotSpot 2.0 Server JVM

My Test subject:

Inbox page consists of an XSP page that checks the
session for an Inbox object.  The Inbox object retrieves
entries from a database, and manipulates objects created
from those entries (sorting, etc).  XSP is followed by
5 stylesheets (one of which is selected dynamically),
and the HTMLSerializer

Test 1 (Simulate 3.456Million hits a day):

Number of Threads: 12
Delay in ms:       300ms
Deviation:         100ms

According to the Cocoon logs, I my processing times
were between 10ms and 150ms.

Test 2 (Simulate 5.76Million hits a day):

Number of Threads: 20
Delay in ms:       300ms
Deviation:         100ms

I hit saturation on my machine, and I got processing
times between 2s and 2.3s.  This is still pretty good,
so I decided to go all out.

Test 3 (Simulate 17.28Million hits a day):

Number of Threads: 20
Delay in ms:       100ms
Deviation:         100ms

The cache seemed to kick in better this time--probably
from running the tests back to back without restarting
Tomcat.  I got processing times between 750ms and 1.5s.

Ladies and gentlemen, I have never seen anything so fast
in my life.  Concidering the complexity involved and the
fact that we have interpreted bytecode makes this feat
even more amazing.

You guys have made me a beleiver in the Caching system.
I thought Cocoon couldn't be much faster than it was,
and boy was I wrong.

To unsubscribe, e-mail:
For additional commands, email:

View raw message