Thinreg performance impact/benchmark

[Matheen]

I have an interesting question. One of our customer is implementing thinapp. They are a mix of VD and physical desktops (mostly physical desktops), all running windows xp pro. 500 desktops in first phase and then expands to few thousand desktops afterwards.

Their implementation plan is to have the packaged thinapp on a network share and use logon script to register their applications using thinreg. Thinreg.exe will be on the base image and the script will be pointing to the packaged applications on a windows share. An normal user will get an average of 10 to 15 applications (small and medium size applications).  All the users will be logging pretty much the same time in the morning. I couldnot find much information about thinreg performance impact on line.

so I want  to know is there any performance benchmark on running thinreg on the above scenario?

Has any one come across similar scneario?

What is the best way of calculating the disk, memory and network impact of running Thinreg for the above scenario?

Regards

Matheen

[Theike]

Hi,

My advise to you would be not to use thinreg for this purpose to prevent experiencing and having to debug quite some undesired behaviour...

For example:

A user gets both office 2007 (registering .doc and .docx) and office 2003 (only registers .doc).

What will happen if the user doubleclicks a .doc file...? This will depend on the order the thinregs will be performed.

So apart of performance, i would advise getting some desktop managing tool. As a professional company we are using the RES Software packages for this, which also help us with the distribution of the applications using the automation manager. (But there are quite some other tools available as well, even for free)...

As for the performance i assume the determining which applications have to be registered wil require just as much time as the real registration. You will have to lookup the users permissions (domain group membership?) and depending on that activate the registration. Doing this for a large group of applications is quite time consuming as well, even if you have only a few applications assigned...

Kind regards,

Michael Baars - Comprehensive ICT Solutions (NL)

(Please mark the question answered if so and reward points to those who helped you)

[pbjork]

We do not really have any benchmark papers on this.. I know we are looking into creating something around deployment in general and I would suspect this to be one topic..

We do have some cool scripts that you can use as inspiration..

http://communities.vmware.com/docs/DOC-13715 needs ThinAppSDK on all clients (http://www.vmware.com/go/thinappsdk).

http://blogs.vmware.com/thinapp/2010/04/simple-thinregexe-login-script.html

It does take some time for thinreg.exe to examine AD permissions. The SDK script does this much faster. Especially secondary logins..

[Matheen]

@thieke - Thanks for your reply. I will have a look at RES software. Can you also let me know of few free tools to manage/monitor application deployment?

[Matheen]

Thanks for the info Pjork. Would be good to have some kind of performance benchmark in future.

If we need to do any testing I will post the results.

[pbjork]

If third party solutions is of interest please have a look at http://www.liquidwarelabs.com/products/profileunity.asp as well..

I am not saying either one is better than the other.. Just want to give you a little more to choose from..