Chat with us, powered by LiveChat

   The Base - 20 Dallam Lane, Warrington - WA2 7NG - United Kingdom      +44 (0) 1925 909650

 

Roaming IE 11 Cookies and History and the Impact on Logon Times

 

One of the main reasons for using Avanite WebData Control, is to bridge the gap between providing users with a consistent user experience in each session and having logon times that are impacted due to roaming large amounts of data that web browsers store in the user profile.  When using a non-persistent environment such as XenApp, where users access a different machine each time they logon the overhead of roaming web data often leads to administrators choosing to not roam this data at all.

In this article we look at how roaming web data for Internet Explorer can impact logon times, and how Avanite WebData Control can be used to mitigate this whilst still providing users with the best possible user experience.

In the following sections we’ll look at the scenario’s covered and the results attained.  The following points should be noted however before continuing to read:

  • Profile Management Settings – The tests are not like for like and cannot be used to form the basis of a comparison between the different Profile Management solutions tested with.  Where possible out of the box configurations have been used and these out of the box configurations provide considerably different functionality.
  • Windows Version – All user sessions were hosted on fully patched Server 2016 Standard Edition virtual machines with 2 virtual cores and 8 GB RAM.
  • User Sessions – All user sessions initiated via RDP.
  • User Accounts – Standard Active Directory domain user accounts were used.
  • SMB versions – All servers used supported SMB dialect 3.0 or above.
  • Contention – As the scenarios were executed in a lab environment all virtual machines only used a single session and had little/no resource contention.
  • Logon Timing Measurement - The Avanite WebData Control Evaluation Reporting website was used to gather logon times.  This measures from the point at which the Windows session is created up to when the user desktop is initialized.
  • Simulated Browsing – Each user used Internet Explorer 11 to browse 25 random websites in each session from the list of the top 5000 websites according to Alexa from early 2018.
  • WebData Control Settings - WebData Control is used in both passive and active modes. Passive mode means it does no management of web data, it merely gathers information. Active mode means it is actively managing web data

Scenario 1 – Ivanti UWM

The first scenario we covered used Ivanti’s User Workspace Management Environment Manager Personalization solution.  We used the Environment Manager 2018.1 SP1 release and the built-in Windows Settings Group called “IE11 Cookies and History (W10, WS2016)”:

The results of the test with WebData Control in passive mode were as follows:

With WebData Control in active mode the results were as follows:

Comparing the 2 sets of data we see the difference in the logon times between active mode and passive mode:

Scenario 2 – Citrix Profile Management without Streaming

The second scenario we covered was using Citrix Profile Management 7.18.  We followed the guidance in the Citrix Support article here - https://support.citrix.com/article/CTX224498, with the following folders added to the mirror folder list:

  • AppData\Local\Microsoft\Windows\INetCookies
  • AppData\Local\Microsoft\Windows\WebCache
  • AppData\Roaming\Microsoft\Windows\Cookies

The “Process Internet Cookie files on logoff” policy was enabled but Profile Streaming was disabled.

The results of the test with WebData Control in passive mode were as follows:

With WebData Control in active mode the results were as follows:

Comparing the 2 sets of data we see the difference in the logon times between active mode and passive mode:

Scenario 3 – Citrix Profile Management with Streaming

The third scenario we covered was using Citrix Profile Management following the guidance in the Citrix Support article here - https://support.citrix.com/article/CTX224498.  Profile streaming was enabled, and the following folders were added to the mirror folder list:

  • AppData\Local\Microsoft\Windows\INetCookies
  • AppData\Local\Microsoft\Windows\WebCache
  • AppData\Roaming\Microsoft\Windows\Cookies

The “Process Internet Cookie files on logoff” policy was also enabled as per the guidance.

The results of the test with WebData Control in passive mode were as follows:

With WebData Control in active mode the results were as follows:

Comparing the 2 sets of data we see the difference in the logon times between active mode and passive mode:

Scenario 4 – Liquidware ProfileUnity

Our 4th scenario used Liquidware ProfileUnity version 6.7.7.6701.  For this test we used the default ”Windows 7, 10 or 2016 as a Desktop and 2016 RDS Full Desktop” template with the “Windows 10 Internet Explorer 11 – 6.7.5” portability settings amended to remove the Process Action After Login setting:

The results of the test with WebData Control in passive mode were as follows;

With WebData Control in active mode the results were as follows:

Comparing the 2 sets of data we see the difference in the logon times between active mode and passive mode:

Scenario 5 – VMWare User Environment Manager

Our last test scenario was using VMWare UEM version 9.4.

Here we enabled the “IE WebCache” built in Windows Common Setting:

The results of the test with WebData Control in passive mode were as follows:

With WebData Control in active mode the results were as follows:

Comparing the 2 sets of data we see the difference in the logon times between active mode and passive mode:

Observations

During the running of the tests the following was noted:

  • When logon takes longer the time taken to logoff also increases.  This was especially visible with the Citrix Profile Management test where profile streaming was enabled as logoff often took several minutes to complete
  • As with all tests there is a slight natural variance observed which means the results are not as linear as one might expect

 

Summary

The following table summarises the results gathered from the scenario’s outlined above:

To summarise, we feel it is enough to say that, over time, as web data accumulates logon times start to be adversely affected.  Avanite WebData Control provides a method via which this “creeping death” can be avoided.

 

 

Peter Jones (Avanite CTO)

Recent Posts

Google Chrome And It's Magic Trick!

Google Chrome – Managing Extensions

Roaming IE 11 Cookies and History and the Impact on Logon Times

End User Computing and Web Browsing

Do You Want to Save Your Password for This Website?

Contents of the WebCache – What you didn’t know.

The Web is the Web