It is tempting to assume SSDs have made server hard drives obsolete. They have not. For the right workload, a nearline HDD is still the most economical storage you can buy. The real skill is knowing which to use where - and most good server builds use both.
Where HDDs still win
Capacity per dollar, by a wide margin. Nearline 7.2K SAS and SATA drives reach very high capacities at a fraction of the cost-per-terabyte of flash. For backups, archives, surveillance footage, media libraries and any cold or warm data, HDDs remain the obvious, cost-effective choice.
Where SSDs win
Latency and IOPS. SSDs are an order of magnitude faster for random access, and they use less power per operation. For databases, virtualization hosts, caching, and anything a user is waiting on, an SSD transforms responsiveness. NVMe SSDs push this even further for the most demanding, latency-sensitive workloads.
Cost, power and endurance
HDDs cost less per TB but more per IOP, and use more power at idle. SSDs cost more per TB but far less per IOP, run cooler and draw less power under load - though you have to size SSD endurance (see DWPD) to the workload. Over a multi-year deployment, the power and performance of SSDs can offset their higher purchase price for hot data.
The hybrid approach most builds use
Tier your storage. Put the OS, databases and hot data on SSD or NVMe; put bulk capacity on HDD. RAID controllers and software-defined storage make tiering straightforward, and you get the best of both: fast where it counts, cheap where it does not.
Bottom line
Capacity-bound workloads -> HDD. Performance-bound workloads -> SSD. Most servers -> tier them, with flash for hot data and spinning disk for bulk.
Need the exact part?
Search our catalog of enterprise drives, memory, power and networking by exact OEM part number - so it seats and runs on the first boot.
Browse the catalog →