Performance issues are among the main challenges that occur during VDI (virtual desktop infrastructure) migration projects. The challenge arises when users experience a drop in performance with their new virtual desktop compared to the performance they had when using a desktop PC.
A drop in performance can be caused by a number of factors, including inadequate compute, memory, or storage capabilities for the applications the user needs to run.
Some of the problems that users can face when there are performance issues with a new virtual desktop include:
- Slow boot-up
- The virtual machine runs slowly during normal operation
- Applications take too long to launch
- The virtual machine becomes unresponsive
Furthermore, these performance issues can be exacerbated by other user experience problems, such as difficulties accessing shared files, connectivity problems, and poor application design.
Whether there are multiple user experience issues or performance is the only problem, the result is usually the same – a drop in productivity and a frustrated base of users.
The Gap Between Virtual Desktop and Physical Desktop Performance
There is a reality that must be accepted by decision-makers and stakeholders at the very outset of a VDI migration. That reality is that there is a performance gap when you compare virtual desktops and the physical computers they are replacing.
In most situations, the physical computer that is being replaced will deliver higher levels of performance. This is because the performance of a virtual desktop comes from its share of the available resources, whereas a physical PC delivers all the performance it has in its box to the user.
There is another reality that is important to highlight too. This is the fact that the benefits of a VDI migration greatly outweigh the potential drop in performance at an individual virtual desktop level. This particularly applies if the VDI migration is properly managed to mitigate the performance challenges
Application Knowledge Is Crucial
The key factor when mitigating performance challenges in a VDI migration project is knowledge. The more knowledge you have before any virtual desktops are rolled out to any users, the better.
Essential information that you need to know includes:
- What applications are needed and who needs them?
- Where are these applications hosted? Are they in the cloud, in a data centre, etc?
- How does the virtual desktop connect to the applications it needs to run?
- Are the applications suitable for a VDI migration?
In other words, it is essential you fully understand the application estate and its topology. This is challenging in an enterprise environment when there are thousands or tens of thousands of users and a similar number of apps. Not only do you need to know what those apps are, but you also need to know if they are suitable for running on a virtual desktop, the level of CPU and memory resources they need, and more.
Automating as many processes as possible is the solution. Here are two examples:
- Access Capture’s Testing functionality assesses the performance of apps after conversion to a modern format such as MSIX or AppVolumes and compares this performance to their original state.
- Access Symphony’s Agent functionality monitors machines and apps before a VDI migration, generating essential performance data that can then be used to ensure users get provisioned and migrated to a virtual desktop that is an appropriate size.
In both the above examples, potential virtual desktop performance issues can be identified and addressed before they are experienced by users.
The Importance of Planning
Planning the VDI infrastructure is another essential part of the process. You must ensure there are enough resources to support the virtual desktops, and you also need to take into account anticipated future needs. However, it’s equally important to avoid over-provisioning resources.
On the infrastructure side of the equation, processor resources, memory resources, storage system disk contention issues, and bandwidth are all important considerations in relation to virtual desktop performance.
There are also further things to look at on the application side of the equation – things that are in addition to determining if apps are VDI migration suitable or not. Examples include:
• Are there more applications within the estate than can be supported by the proposed infrastructure?
• Are any of the applications suffering from memory leaks?
• Are there apps that are putting undue strain on CPU or memory resources?
In relation to the last point above, one common example is antivirus software and how it is configured to work in a virtual environment. When we manage VDI migration projects at Access IT Automation, we take this into account to ensure antivirus software is optimised for the virtual desktops.
In some situations, it is also worth checking the version of the virtual desktop product you are using as well as any configuration tools. For example, if you are using VMware, checking the virtual desktop is running the correct VMware product and has VMware Tools installed.
Sometimes, Virtual Desktops Are Simply Not Possible
Throughout the investigation and planning stages of a VDI migration project, it is likely that you will find users and/or apps where migration to virtual desktops simply isn’t possible. An app might not be suitable, for example. Some users might also need to use a combination of apps in such a way that makes a virtual desktop unviable.
The fact that there will be some users that you can’t migrate to a virtual desktop needs to be understood and factored into your VDI migration plans from the start.
The Right Level of Performance for the Situation
This blog has largely focused on the performance challenges that can arise during VDI migrations.
In closing, however, it is important to also highlight that with good planning and the rapid automated insight provided by tools like AppScanCapture Testing and Symphony Agent, the vast majority of users will experience no discernible performance degradation when they are migrated to a virtual desktop.
For that small group of users where performance may present problems, early identification and clearly defined mitigation measures are the solution.