Windows 2003 Application Performance Suggestions

Microsoft Windows 2003 Server has a feature called “Application Pooling and Recycling” not available in Windows XP or Windows 2000.  This feature allows you to enhance application performance.

Application Pooling allows you to increase the stability and reliability of single threaded applications by generating multiple processes which effectively act as one application. This creates redundancy for failed instances because if one process fails the others are unaffected and available to service activations.

Application Recycling further increases reliability by periodically recycling individual instances of running processes which in turn frees the memory it was using. The old process is not shut down until all external references to objects in the process are freed or the expiration is reached, this ensures that the client application is never interrupted by the recycling process. This also ensures that memory and other resources are maintained within acceptable limits.

For more information on Application Recycling, see:

http://www.microsoft.com/technet/prodtechnol/WindowsServer2003/Library/IIS/24e3c22e-79a9-4f07-a407-dbd0e7f35432.mspx

Iron Speed recommends that you set the Maximum Used Memory setting in the Application Recycling to about 75% of the physical memory available in your server.  If your server is running multiple applications and other server software, you may want to reduce this to about 50% of the available physical memory.

To change the Maximum Used Memory, start Internet Information Services (IIS) Manager.

Step 1:  In IIS Manager, expand the local computer, expand the Application Pools folder, right-click the DefaultAppPool or the Application Name, and then click Properties.

Step 2:  Click the Recycling tab.  Check the Maximum Used Memory (in megabytes) checkbox and set the site is megabytes (1 gigabyte is 1,000 megabytes).  When the memory consumed by IIS reaches this limit, the process is recycled.  You can change other settings if necessary based on the server load and the number of users accessing the application.

 

Step 3:  Click Apply, and then click OK.

See Also

Application Scalability and Performance