Unified Service Desk (USD) version 3.2 contains a new options called “InternetExplorerPooling”. But what impact will this option make?
To enable IE Pooling you simply create a UII option called “InternetExplorerPooling” with a value of true.
The Microsoft release notes explain that we can experience enhanced performance of CRM entity page loading in Unified Service Desk with the Internet Explorer Pooling feature.
Unified Service Desk always maintains a pool of Internet Explorer instances for hosted controls to use. Opening a hosted control using a pooled Internet Explorer instance enhances the performance of the Unified Service Desk.
- The Internet Explorer pooling feature supports only CRM entity pages hosted in CRM page hosted control.
- When you enable the pooling feature and open a CRM page hosted control, you can see in Task Manager that Unified Service Desk creates a number of Internet Explorer process instances for hosted controls to use.
- The performance of the Internet Explorer pooling feature is dependent on the resources available on the client computer.
So …. Assuming you are using IE Process hosted controls your CRM pages will load faster. “Dependent on the resources available”.
Great. BUT …..
Microsoft don’t make any comment on the possible performance gains. So I wanted to quantify an actual real-world example. Obviously every system will be different and the benefits you gain will vary. But I wanted to test in my USD instance to give a sense of the possible performance gains. I wanted to know if the performance enhancements are significant or marginal.
In my test I started a session and then closed it again. My config is pretty complicated involving loading a contact and account record. Plus creating a phone call and opening that. I also do loads of other “stuff”! Additionally (importantly), when my sessions close I write out some audit information. This audit information contains the duration of my session. So I could use that to compare my test results.
To ensure my test was as accurate I repeated it 40 times. 20 without the new InternetExplorerPooling option and 20 with it. I then took an average timing.
Initially each test cycle took an average of 38 seconds. After adding the new parameter my average test cycle reduced to 27 seconds. Meaning I achieved a 28% speed improvement.
With this kind of performance gain I strongly recommend you try this, ASAP! Obviously the number of CRM Pages being loaded will impact your results. As will the resources available on the local client. But my results suggest you should see a tangible improvement. Enjoy.
2 thoughts on “USD – InternetExplorerPooling”
I was able to test this feature out in our CRM organization and while it worked as described with a single session tab, it killed the USD performance if I tried having 2 session tabs or multiple pages in a CRM hosted control.
Trying to switch between session tabs or pages would result in USD becoming unresponsive for 15-20 seconds. Closing a session would also take a very long time.
There were no errors listed in the log files, so I’m not sure why this feature doesn’t work for me.
Have you heard of anyone else having issues with it?
Your experience isn’t the same as mine or the expected behaviour.
It might be worth you trying the best practice analyser just released in v3.3. That might help point out some issues.
Failing that I suggest maybe raising a support ticket with Microsoft as they might be able to give you a view.