In order to properly maintain virtual desktop infrastructure (VDI), how important is scaling? After all, there are many questions which arise related to scaling and VDI.
In fact, Mark Plettenberg, product manager at Login VSI, said many of their customers often ask about scaling tests. He said the question often comes as follows: "If I want to find out the number of servers I need, I simply test how many users I can put on one server then multiply by the number of users I need to support. Right?"
The answer is no. He explains that is because the environment is different when testing on one vs. two machines.
He further explains in a blog post it is better to start with one machine when doing performance tests. Then, increase to full scale. “I highly recommended to test at scale so you know exactly what level of performance to expect,” Plettenberg said.
The reason why testing should be done in this way is because of storage systems, Plettenberg said. “It’s very likely that you are using shared storage. Meaning, at some point you will be out of storage resources. And without storage optimization solutions, this might be sooner than you like,” he added.
It is interesting too that when considering complex environments there are more shared components, such as with: broker/load balancers; monitoring; storage; power; networking; cooling; antivirus; backend databases; and file servers.
“The only way you can truly validate that all of the above are working correctly together is by testing at full scale,” he explained. In addition, it is important to note that after using just a few servers, scaling becomes non-linear, another wrinkle.
Testing at capacity can also help users answer such questions as: Is an environment stable and performing at capacity? Do backend systems handle the load? Will an environment handle logon-storms? What if a data center, server or rack fails?
In addition, testing at full-scale shows when environments will not work. In one example, he cited, power was pulled from the grid during a Login VSI test. In another instance, a storage device was misconfigured and used a limited amount of CPU.
Another article with related advice relates how Alastair Cooke recommends in Tech Target adding of “more virtualization hosts when you need to accommodate more users.”
A VDI pilot can help someone find the right number of users than can be used on a single host or a cluster of hosts, he said. Then, find out where bottlenecks and single points of failure can occur. “These are unavoidable, so understanding where they lie will enable operational practices that minimize their impact,” Cooke said.
Edited by Rory J. Thompson