In my understanding Thin Client Analysis Module (TCAM) is able monitor the performance only after RDP session is established on RDP Server. With that understanding the response time for end user will be measured between presentation layer and back-end system. Is there any way response time of transaction can be monitored from Thin Client/User Desktop?
"Is there any way response time of transaction can be monitored from Thin Client/User Desktop? " short answer is no, not if you are looking at the RDP data stream only.
This is beacuse the RDP protocol is (mostly) screen updates and don't necessarily relate to what's going on in the back-end.
The only way to monitor the end-user performance would be to combine this with synthetic enterprise. The advantage is that these robots can be a very good reference to interprete the DC-RUM data. And the combination with DC-RUM will allow you to not script every single user transaction but focus on 5 or 6 key ones. We have customers like ITERGO who even presented this on Perform this year.
The Citrix/WTS (presentation) tier belongs to the Network Tiers section because it is treated as an application delivery channel. For Citrix and Windows Terminal Services, the key metrics are network performance, server RTT, or server loss rate, realized bandwidth; not operations or operation time as it is for data center tiers (back-end).
Notice that it's not recommended to only use TCAM but also use analyze the traffic in front (arriving to) of the Citrix and at the back (traffic between the back-en and Citrix).
Yes - as long as you don't mix up what the users "see" and what the "do". RDP/ICA protocol has nothing to do with what they run in their screen and most users (except engineers) don't understand the difference of a f.ex. a published application having problem or the RDP session having problems.
Thank you for this post
Do you know, how are the Citrix KPI when there is a performance issue only on Citrix?
Are there thresholds (on CPU, memory metric etc) to qualify the Citrix performance ? and the EUE for this tier?
Sorry I missed responding on this in June. The performance metrics (KPI's) of Citrix is well known. According to Citrix, a conversation has to have a better RTT than 125 ms, a losss rate less thatn 0,1% and a realized bandwidth of more than 26 kbps (in reality even better)
There is no such thing as a EUE for Citrix as it is non transactional. The Eue is for the app that runs IN citrix, it has nothing to do with the Citrix itself.
These 2 are not related to each other even though they deliver something together, but for the end user this is never apparent (clear).
TCAM is for 2 things (I probably forgot something):
So I wouldn't go without that - and don't miss to read Chris PPT (I'll send you the link in an email)
TCAM is absolutely necessary for XENAPP (both desktop as published APP) and WTS. The reason is that you have many users sharing one Server Platform. In order to understand which CITRIX/WTS user used what applications in the back-end you can only do this with the mapping that Ulf is refering too.
For XenDesktop there is no TCAM agent.
1) Because the XENDESKTOP environment is only 1 user - 1 desktop relationship, so there is no mapping issue. Whatever was done by this one desktop ip is immediately related to this one user. The ICA decode will give us the userid.
2) The resources are slightly less important because it is not shared with many users <-> XENAPP, and no we have no immediate plans for creating an agent.
I hope this clarifies the TCAM and its usage a bit better.
You mentioned for XenDesktop, ICA decode will give us userid. Is it done by default or it needs to be configured?
"For XenDesktop there is no TCAM agent.
1) Because the XENDESKTOP environment is only 1 user - 1 desktop relationship, so there is no mapping issue. Whatever was done by this one desktop ip is immediately related to this one user. The ICA decode will give us the userid."