A virtual desktop infrastructure (VDI) can help reduce operating costs by as much as 90 percent by increasing resource consolidation and bolstering performance while providing an exceptional user experience.
While this is all excellent news, there’s a lot that needs to happen to ensure a successful launch of a virtual desktop environment, and that includes performance testing. Without the proper care to assure your VDI can handle many users’ workload all logging on and using your infrastructure at once, your VDI may do more harm to your operations than it does good.
It’s kind of like when car manufacturers test every safety feature and engine component of a new car prototype before starting mass production; they need to be sure that the car is sturdy enough to withstand inclement weather and harsh driving conditions, all while protecting the passengers even in the event of a crash.
VDI performance testing is similar. This testing is designed to ensure that your virtual infrastructure works no matter what. And yet, in one survey about VDIs, 44 percent of respondents said they never did any kind of performance testing before launching their VDI.
What can happen if you don’t test your VDI before you launch it? These real-world scenarios serve as cautionary tales:
- A VDI that failed with just over 4,000 users because the virtual desktop consumed too much power
- Several instances of VDI environments that relied on provisioning servers that did not have a large enough capacity
- VDI environments where only 1 out of 20 total CPUs was used, which was a waste of resources
- Updating from one version of an operating system to the newer version and losing 20 percent of the users in the process
Your virtual desktop infrastructure can have many benefits. But only if it’s “road tested” first to ensure it won’t be a bumpy ride.
All About the User Experience
Having a virtual desktop infrastructure only matters if it works! You may have a band new VDI, but it won’t matter if the user experience isn’t up to par. Your end-users aren’t worried about whether or not you just implemented the latest and greatest VDI—their only concern is that it performs as well or better than it ever has before.
What happens when a network is oversaturated with traffic or if the backend servers for a virtual desktop infrastructure become overloaded? Application performance suffers, and the user experience leaves a lot to be desired. And in some extreme cases? Users might notice a significant delay. For example, there might be a second or two of delay before the time they press a key on their keyboard and the time the character appears on the screen.
Before launching, a virtual desktop environment may work reasonably well when just a few users are logged on. Still, in actual applications, problems can arise if the infrastructure isn’t thoroughly tested first. This is where VDI performance testing comes into play.
Common Mistakes that Accompany VDI Performance Testing
Anywhere you read about VDI performance testing, the focus is always on the end-user experience. Through performance testing, you can learn more about how to fine-tune your system to ensure the best possible outcome.
But sometimes, performance testing doesn’t necessarily go according to plan. When errors are made, you may miss out on important places to improve before the virtual environment launches.
Here are a few common mistakes that developers make when deploying and testing a new VDI:
- Not testing at all, especially before having real users log onto the VDI
- Starting with a full-scale test instead of testing for a few users first, then testing one server and gradually building up to a full-scale test
- Only testing with your applications and workloads instead of testing other applications and operations to see how they work
- Assuming that tests will scale linearly and only testing with one server instead of scaling things up
- Disregarding “stuck sessions” when things go wrong. Too many stuck sessions could indicate the sign of a larger problem with the configuration
- Not continuing to regularly test a VDI’s performance after it’s been deployed
- Over- or under-provisioning resources
Helpful Hints to Test Your Virtual Environment
What kinds of information should you include in your performance test? Like how car manufacturers test all types of road conditions at various speeds to test performance, you need to be aware of multiple factors as you test your VDI.
There are a few considerations to build into your test to truly understand how your virtual environment will perform in the real world. These include:
- The resources required by each user and your overall storage and memory requirements
- Desktop login time to see how and when resource shortages happen during boot storms. You should create a “boot storm,” or a test of many users logging on at once to see how your VDI holds up and what load times are like.
- Comparing application boot time and overall performance within a VDI to performance on a PC that’s running the same applications and operations in a traditional desktop environment
Unless you’ve tested a virtual desktop infrastructure before, it can be hard to know what to look for; what benchmarks to set to know if your infrastructure will serve you the way you want it. A few considerations your VDI testing checklist might be:
- Pre-determining what a “successful” test will look like
- Defining the parameters of your testing to evaluate basic functionality, which should include:
- If you have enough resources like CPUs and memory to support your VDI
- The network connection is strong enough to support your VDI
- Run user acceptance testing and having several real users test the VDI
- Evaluating your tests and making necessary adjustments
When you perform a deep dive into the functionality of your VDI, you can discover what works and what doesn’t and make revisions before launching your VDI to ensure successful implementation at the moment of truth.
The VIOLIN Systems Advantage
Suppose a successful virtual desktop infrastructure is reliant upon providing a great end-user experience. In that case, the key to this success is powering your VDI from a storage platform that’s up to the task.
Whether implementing a simple virtualization infrastructure or powering from a private or public cloud setting, VIOLIN’s QV2020 all-flash array offers a secure environment and consistent performance even during boot storms and peak loads.
And the best part? The VIOLIN QV2020 supports a state-of-the-art user experience and top-notch security at a fraction of traditional desktop infrastructure costs. VIOLIN’s all-flash arrays can support thousands of virtual desktops without latencies and a dip in performance.
If you’re considering creating virtual desktop infrastructure, one of the most important decisions you can make is your storage platform. VIOLIN’s QV2020 all-flash array is affordable, secure, scalable, and will ensure consistent performance no matter what. To learn more about how our all-flash systems can support your VDI, contact us today!