When IT planners are asked to list their top challenges with virtual desktop infrastructure (VDI) they often come up with a variety of answers, like gaining user acceptance, delivering an experience similar to that of a stand-alone desktop or driving down the cost per desktop. But in practice it always comes down to performance. A fast, efficient environment will allow you to deliver a faster-than-desktop experience and drive higher user acceptance. It will also allow you to reduce the number of physical servers required to host the VDI environment and reduce the amount of storage (and number of storage systems) you need to support that environment’s data.
Watch the On Demand Webinar "How to Overcome the Four Obstacles to VDI Love"
You need more than flash
The default answer it seems is to throw flash, any flash, at the problem. There’s no question that flash can improve performance of the VDI environment, but selecting the right type of flash storage and making sure it produces the maximum performance gain is also key.
The VDI environment’s continued demand for performance is unlike any other in the IT arena, even databases where once you reach a certain amount of IOPS, any additional performance is wasted. Not in VDI. Any additional IOPS that a storage system has available can be consumed by increasing the number of virtual desktop instances per physical server, as long as the performance of each virtual desktop does not degrade as density increases. The fewer physical servers that can be deployed the lower the cost per desktop.
There are plenty of processes that consume IOPS outside of the virtual desktop instance itself. Most of these processes are designed to reduce the storage capacity requirements of the VDI environment. Features like thin provisioning, golden masters, clones, deduplication and compression have significantly reduced the storage capacity requirements of a typical VDI implementation. In fact, we rarely hear managing capacity as a top concern any more.
But what is the impact of these features? In many cases they lower the overall performance efficiency of the flash storage system itself. Yes, flash made the status quo perform better, but by itself, did not allow for a significant step forward in the number of virtual machines per host. This was because flash subsystems are burdened with managing all the extra write I/O that these efficiency processes create.
Capacity-saving features need to be modernized so that they can perform their originally intended functions without dragging down the performance of the all-flash array. If this can be accomplished, then the number of desktops per host can be increased to the point that the cost per desktop is lowered while users fall in love with their virtual desktop.
New technologies are indeed enabling VDI systems to provide ‘capacity efficiency’ without affecting ‘performance efficiency’ – and increasing virtual desktop density in the process. To learn more about how these technologies can make users fall in love with their virtual desktops, watch the on demand webinar “How to Overcome the Four Obstacles to VDI Love“. When you register you will have access to our exclusive white paper “The Three Ps to a Successful VDI Deployment – Persistence, Price, Performance”.