How to Roll Out Citrix XenDesktop Globally, Without a Hitch

In addition to Pieter Van Heck, Alasdair Patton (Delivery Architect for Dynatrace) and Mark Revill (Field Technical Services, Dynatrace) also contributed to the authoring of this post. 

Windows XP is living on borrowed time: Microsoft officially ended extended support for it on 8 April 2014. It was inevitable, but now many companies and organisations across the world face the unenviable task of removing XP from their organisations without disrupting regular business activities.

One of our customers, a bank with multinational operations, took this opportunity to implement XenDesktop, Citrix’ Virtual Desktop Integration (VDI) package, and make managing their employees’ computers cheaper and easier going forward in the future.

The technical team within the bank responsible for the rollout set themselves up for success by doing two things: first of all, they tested their implementation on small groups of employees before the main rollout. Initially they deployed XenDesktop to an offshore office where everything worked fine. Then they deployed it to a bank branch near their headquarters, and there were a few user complaints. The VDI introduced an extra stage in delivering applications to end-users, and at this point the technical team realised they didn’t have visibility into its performance.

Even at this early stage, the VDI architecture was complex, consisting of hundreds of Citrix severs, Web servers and other back-end systems. Manual performance diagnosis would be time-consuming, disruptive and difficult.

Regaining Visibility

Here’s where the second success factor comes in: the bank added Dynatrace APM Citrix monitoring to their existing APM setup. With this, they restored end-to-end visibility into their systems’ health and learned how it was affecting user experience.

Below is a screenshot showing the APM Portal (Picture 1), which the bank’s technical team uses to monitors and field alerts on the performance and availability of:

  • The business application delivery chain (VDI Deployment) from the virtual desktop instance through application load balancer, front end servers, middleware, to the back-end tiers;
  • The Web interface that the VDI users (Desktop Launch) contact to log on to their desktop instance;
  • The desktop-to-display network links (VDI Channels), including breakdown into types of end user activities with the desktop (screen, printing, multimedia etc.).


Picture 1

All three are needed for quick fault domain isolation. Even with the added complexity of the VDI, it’s possible to isolate a performance problem within minutes, as compared to the weeks required for manual diagnosis.

The bank’s technical team immediately saw the benefit of this visibility when encountering a strange problem: some users complained that they couldn’t login and work on their machines. Their screens were just hanging, and not doing anything. But users who were already logged in could continue working.

The users couldn’t give a detailed technical report of their problem—all they saw was that Citrix wasn’t working for them!

When the bank’s technical teams looked at the issue (Picture 1) they could immediately see that the issue was related to the Web interface (Desktop Launch). They drilled down into the login process, and saw that there was a problem with the Web interface, as shown in the screenshot below (Picture 2):


Picture 2

The number of HTTP server errors had dramatically increased, preventing users from logging in over the VDI’s Web interface. Other components of the system were unaffected, which is why those who were already logged in could still use their machines.


Network performance is important to performance in Citrix environments; this much, the team already knew. However, with the benefit of full visibility into their environment, the technical team enriched their understanding of and learned which metrics lead to good performance from the end-user perspective—in the case of XenDesktop, network performance characteristics like realized bandwidth are key, for instance.

Furthermore, as every component of their application delivery chain is visible to them, the team can isolate and solve any issues within a matter of minutes. In the case of XenDesktop, it’s more than the display-to-the-server network link: it includes authentication and provisioning services that have to perform in order to instantiate the desktop instance that the user could log on to.

Ultimately, the strategy of having complete monitoring in place before a full rollout has a big payoff: the way is cleared to implement VDI globally in their organisation—without any hitches.

Pieter Van Heck, is the Sales Engineering Manager for EMEA, in his role he focusses on the Data Center Real User Monitoring Solution. Pieter has been working in the IT for over 15 years, where started his career on the Mainframe. The last decade of his career Pieter has worked on the different set of technologies around End-User Experience, therefore has a unique experience around the value, usage and implementation of end user monitoring solutions and how the solution market has evolved.