Third-party Performance and Load Benchmarks for SugarCRM

sugarcrmdevelopers —  March 4, 2008 — Leave a comment

A couple recent performance benchmark reports have been written that use SugarCRM as the test app. Both reports are focused on benchmarking specific environments and only use SugarCRM as the test app, but they do give you a good idea of the scale possible with SugarCRM.

  • Scaling SugarCRM with MySQL on Sun’s Coolthreads server by Satish Vanga at Sun Microsystems. Focused on showing the scalability of SugarCRM on Sun hardware, this report shows the large scalability possible (700 concurrent users!) with SugarCRM and MySQL on the Sun Coolthreads server.

    From the report…
    Conclusion
    In our tests we found that the Sun’s Coolthreads server can scale to large number of concurrent users on a single server running both SugarCRM and MySQL. There is not much performance impact if the MySQL is run inside Solaris Containers vs running it in the global zone along with SugarCRM. At 700 concurrent user load we only saw 15% of the network bandwidth used and still left with enough cpu cycles. Sun’s Coolthreads server does scale well with the SAMP stack based applications more specifically SugarCRM and MySQL database.

  • Load Testing SugarCRM in a Virtual Machine done by http://www.webperformanceinc.com. Focused on showing the performance differences between running a web app in a virtual machine vs a dedicated physical server, this report doesn’t look to tune a system for maximum performance like the Sun article but rather to take a basic SugarCRM installation and compare the load and performance results on a VM vs a physical machine. Note that the user counts below are concurrent users and not named users.

    From the report…
    Conclusion
    Under the conditions tested, the reference application (SugarCRM) showed a 14% decrease in total user capacity when running in a virtual machine (decreasing from 35 users to 30). At the peak capacity of the virtual machine (30 users), the average page duration was 1.57 seconds, whereas the average page duration of the physical machine at the same load level was 0.97 seconds. Compared at that level, the average page duration for the virtual machine increased 61% over the same load on the physical machine.Since most applications only run occasionally at full capacity, it is worth also analyzing the performance of the system at a lower level of load. At 21 users (60% of the full capacity of the physical machine, 70% of virtualized capacity), the average page duration was 0.84 seconds for the physical machine and 1.00 seconds for the virtual machine – an increase of only 19%. For our reference application, this difference (0.16 seconds) is unlikely to be noticed by the average user.

Both reports have really useful information if you’re interested in the details of SugarCRM scalability and performance. It’s great to see SugarCRM becoming a default test app for these different scenarios. Enjoy!

No Comments

Be the first to start the conversation!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s