Should I use an all-flash or hybrid array for my virtual desktop environment?

The first factor to consider when deciding to use an all-flash array or a hybrid array in a virtual desktop environment is cost. All-flash arrays provide a cost advantage when calculating IOPS per dollar. A hybrid array, which mixes flash devices and traditional hard disk drives (HDDs), does better when calculating raw storage capacity per dollar. Things get a little fuzzy when calculating cost per usable gigabyte (GB) of storage; many all-flash arrays do an excellent job of compression and deduplication, so they may be cost-competitive with hybrid arrays when usable capacity is taken into account.

In the early days of all-flash arrays, the technology had limited use cases. As the technology matured and the cost of flash came down, the use cases expanded and became viable for virtual desktop environment deployments. The major advantage of using an all-flash array is its speed, which is the result of decreased latency and increased IOPS. Speed is a valuable factor in mitigating boot storms, but it can also enable inline deduplication, compression and thin provisioning. Virtual desktops are very homogeneous and compressible, and all-flash arrays are very good at compression and deduplication. Some all-flash arrays claim up to 80% reduction in storage space. When all-flash arrays are used in a virtual desktop environment, they can become cost-competitive with hybrid arrays.

When all-flash arrays are used in a virtual desktop environment, they can become cost-competitive with hybrid arrays.

Hybrid arrays, in theory, allow the user to enjoy the speed benefits of flash with the cost benefits of HDDs. While it’s not possible for all data to be on flash, the tradeoff is that cold or infrequent data can be stored on less-expensive HDDs. If virtual desktop users need to frequently access cold data, they may become frustrated with the amount of time it takes to pull this data from HDDs.