That's good information but I'm afraid it can be misleading.  Below is my experience 
with Cocoon server performance.

The Cocoon framework provides the basis of a Fortune 500's intranet where I consult.  
They have almost 6000 employees and in addition to the obligatory "brochureware" 
content there are several applications where employees conduct business.  The 
thousands of information pages are actually pulled through aggregation pipelines where 
the "content pane" is wrapped with typically 5-6 other fragments.  (Then the links 
have to be rewritten and various other transmogrification steps per the business.) 
There are a few database calls scattered throughout and online forms but primarily 
this is a site that is aggregating static XMLs (which are managed via a separate 
content management product). 

Unfortunately, the original pipelines are too long.  There can be upwards of 15 XSL 
transformations taking place to produce one page.

Also, there are many large BLOBs (PDFs, XLS, DOC, etc.) that Cocoon seems to have 
difficulty serving up.  Cocoon would prempt its own response and return a 0 byte file. 
 So the business has worked to pull all non-essential requests out of the Coccon 
space.  All BLOB links (including images, CSS, and js) are now handled by Apache HTTPd 
because it makes sense.

So what are the metrics?

USAGE PER DAY
--------------
 4,000 visits 
60,000 pages
95,000 hits

THE SERVER
--------------
Dell 2550
Dual P4 2.0GHz
1.0 GB RAM
Tomcat 4.1.27
Apache 2.0.47

Not earth-shaking but enough to make the server (and Corp Management!) notice.  Plus, 
our usage is "scheduled", with most workers coming in around the same time, checking 
the online lunch menu around noon and going home en masse.   Everything runs well 
within limits but I'm sure that things could be tuned better.

Sorry for the verbose message but I wanted to give my real-life example.  I think that 
Cocoon is a wonderful framework that benefits from an active developer community 
(READ: this list!).  As with all Apache projects, the documentation is abundant enough 
but sometimes I do struggle to find information.  Hence my questions below.

--What are the best practices regarding pipeline length?  It appears that threads are 
being spawned to facilitate these transformations.    

--Where can I go for some guidance on practical XSLT would be beneficial for not only 
myself but the entire Cocoon community? 

--Finally, what is the definitive answer regarding aggregation invalidating the 
caching mechanism?  (I read that sometime ago on this list.)

Peace,

Luke.....









-----Original Message-----
From:   [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent:   Fri 10/24/2003 10:20 AM
To:     [EMAIL PROTECTED]; Sergey N. Fefilatyev
Cc:     
Subject:        Re: cocoon sites


they are 13 live sites in our production server till now

p4 2.4
1G RAM
redhat
cocoon2.0.4 on tomcat

and the machine is allmost sleeping all the time 


--stavros
On Fri, 24 Oct 2003, Sergey N. Fefilatyev wrote:

> Hello fellow developers!
> Could you tell me , how many sites(the maximum) is possible to server
> with cocoon?( i mean with one instance)
> 
>   
> 
> 


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]







<<winmail.dat>>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to