ADF Applications Hardware Sizing?

[Shay]

I was wondering if members of the group would be willing to share the
system configuration of their production ADF applications.

We sometime get questions from organizations who are starting out with
ADF about hardware sizing for their application - this is in terms of
how much memory per user/ how many users per CPU etc..
The answer to this of course is "it depends" - since there are so many
factors that can influence the behavior of an application there is no
hard number or rule of thumb we can provide.
It would be interesting though to see if we can get to some average
numbers by analyzing the data for current implementations.

I guess the information that would help would be something like:

* Complexity of pages: simple(basic pages), medium(dynamic UI - tabs,
accordions etc), complex (regions/taskflows)
* How many users (maximum and concurrent users)
* Server OS
* Number of CPUs
* Estimated memory per user

[Chris Muir]

I'm a bit surprised with 300 odd members nobody has bothered to reply
to Shay. Can I remind members of the group, which is becoming my
quarterly nag, that the success of this group is driven by you the
members. If nobody participates, the group might as well not exist.
So please take time out to give Shay a response.

In response to Shay, for my current client's production WLS
application, we're not pushing the envelope at all, with over the day
just 60-120 users, and on average 1 concurrent user at any one time
(I'm not sure the terms "1" and "concurrent users" can be used
together). However we're in the process of writing and delivering
more subsystems which will be deployed on the one WLS box which will
stretch it somemore. While Oracle has delivered a framework for
writing large scale applications, we're successfully using it to build
small specialized applications too.

Meanwhile Shay to reverse your question, is there any documentation
that exists in helping people to estimate the hardware requirements?
In turn any documentation for sizing the ADF BC application and
connection pools correctly beyond the whitepapers that describe (dare
I say simply) the AM pool features?

Cheers,

CM.

2009/12/9 Shay <shay.sh...@gmail.com>:

> --
> You received this message because you are subscribed to the ADF Enterprise Methodology Group (http://groups.google.com/group/adf-methodology). To unsubscribe send email to adf-methodolo...@googlegroups.com



--
Chris Muir

Oracle Systems Consultant & Trainer
SAGE Computing Services Pty Ltd
http://www.sagecomputing.com.au

Oracle ACE Director
http://www.oracle.com/technology/community/oracle_ace/index.html

Email: chris...@sagecomputing.com.au
Blog: http://one-size-doesnt-fit-all.blogspot.com/
Mobile/Cell: 043-828-6421
International: +61-43-828-6421
Gmail chat: chris...@gmail.com
Twitter: twitter.com/chriscmuir
LinkedIn: http://www.linkedin.com/in/chriscmuir

[amir hosein khanof]

Dear Shay,Chris,

 

Regarding the hardware information and our system scale pls find the below.

I have attach a image regarding this which shows the cpu and the RAM.

4 cpu 3.4 GHz and 16 gig RAM in windows 64 bit server 2003 service pack 2.

 

we have 4 deployed system each with about 10 page and 15 task-flow and around 50-70 business component objects

the number of user are around 800 numbers and the number of cuncurrent user are  around 50-60 users and in the peak we have around 100-120 concurrent users, the RAM usage is from 4 gig to 12 gig and after 20-30 days which server is up we need to restat the weblogic in order to free-up the memory which I do not know why when the number of user is downsizing for example from 120 to 40 the memory is not reducing and even by doing the garbage collection only 1 gig is freed!

 

regarding the connection pool we are using 2 connection pools one with Maximum Capacity:300 and the other with 80.

our deployed release is 11.1.1.0.2 we are facing many bug's in production environment and now we are deploying to 11.1.1.2.0 hoping solve our problems.

 

Regards,

Amir

[Andrejus Baranovskis]

In my project I was presenting on OOW'09: http://andrejusb.blogspot.com/2009/10/slides-oracle-adf-11g-production.html we are using 2 hardware machines, each with 8 GB RAM and 4 GHz Dual Core. We have configured WebLogic cluster on RedHat Enterprise 5, however without replication at the moment. There is hardware load balancer as well, to route work load between cluster members. Regarding JDBC data source we have configured 5000 as maximum capacity, with 300 seconds period clearing inactive connections.

Deployment in another project we are doing is based on a set of small apps, there configuration is pretty simple - 6 GB RAM, 4 GHz Dual Core, RedHat Enterprise 5.

Regards,

Andrejus

2009/12/15 Chris Muir <chris...@gmail.com>

--
Oracle ACE Director

My Blog - http://andrejusb.blogspot.com/
My JDev/ADF Samples list - http://andrejusb-samples.blogspot.com/

[Marko Mitic]

In my project (Primary Health Care - GP application) we are using one machine as App server and one machine as DB server. 

App server and DB server configuration are 4GB RAM, CPU Intel Xeon 4 core 2.2 GHz and 4GB RAM, CPU 2.5 Dual Core respectful. 

DB server is Oracle 10g Standard Edition One on SUSE 10. Our DB storing administrative and medical data for 120000 patients. We have about 2000 doctor-patient contacts per days.

Application was developed on ADF  11g TP3 (EJB 3.0) and not yet migrate to WLS and one of production release of ADF 11g. Application have about 20 pages/taskflows for doctor role and about 15 pages for nurse role.  We have about 250 users with 20-30 concurrent users.

Because of Technical Preview and a lot of bugs in framework we are reboot OC4j every night.

Regards

Marko 

Marko Mitic

Bulevar Kralja Aleksandra 249/62

11000 Belgrade

Serbia

+381 63 7737328

[BradW]

Hi Andrejus.  How many concurrent users did this configuration support?

[BradW]

Hi Adrejus, how many concurrent users was this actually supporting?

On Wednesday, 16 December 2009 04:03:34 UTC-8, Andrejus Baranovskis wrote:

[Jeevan G Joseph]

My approach is generally to start between a 5-8 Gb heap for each managed server and tune it based on analysis from visual VM(hotspot) or JRMC(jrockit). The stable live set gives a good indication of how much memory you will need. I also try to ensure 1-2 cores for the JVM , especially if I'm using virtualization.

Once the application is fairly feature complete, I run a single user test going through as much of the use cases as possible . Then I take a heap dump from the VM , and analyze it to see the exact size of my session object, which would be an approximation of how much memory a user would consume when the session is active. Extrapolating this to about 90% of the heap size gives a rough idea about how many user sessions I can support with that sort of sizing, without running out of memory or gc'ing excessively. 

If I need to scale, I try to scale the number of managed servers in the cluster than just throwing more memory or CPU cores at it (this of course depends on how much infrastructure I have available to me )

This is by no means perfect, but usually serves as a good starting point .

Thanks,

Jeevan

--

--
You received this message because you are subscribed to the ADF Enterprise Methodology Group (http://groups.google.com/group/adf-methodology). To unsubscribe send email to adf-methodolo...@googlegroups.com
 

All content to the ADF EMG lies under the Creative Commons Attribution 3.0 Unported License (http://creativecommons.org/licenses/by/3.0/). Any content sourced must be attributed back to the ADF EMG with a link to the Google Group (http://groups.google.com/group/adf-methodology).
 
---
You received this message because you are subscribed to the Google Groups "ADF Enterprise Methodology Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to adf-methodolo...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
Sent from Gmail Mobile