Video Production Storage: Tiers, RAID, Network and Archive
Video production storage fails in a specific way: not by running out of space, but by not delivering frames fast enough to keep an editor working. This covers what production storage actually needs, and why the answer is different from general file storage.
What does video editing storage actually need?
Sustained throughput, more than anything else.
An editor scrubbing through high-resolution footage needs the storage to deliver a continuous stream at the footage’s data rate, without stalling. Several editors working at once multiply that. Rendering and export add sustained writes on top.
Three numbers to size against.
The data rate of your footage format. Compressed proxies need little; high-bitrate camera formats and uncompressed intermediate formats need a great deal.
Concurrent streams. How many editors, and how many streams each timeline pulls at once.
Working set size. How much footage is active on a project at one time, which decides how much fast storage you need against how much can sit on slower capacity.
DAS, NAS or SAN for video editing?
Depends on how many people share the footage.
Direct-attached (DAS) for a single workstation. A fast enclosure over Thunderbolt or SAS delivers the highest throughput to one machine with the least complexity. Our storage architecture guide covers the categories.
Network-attached (NAS) for a small team. Shared over 10GbE or faster, with enough drives to serve several streams. The network becomes the bottleneck before the drives do, so the network must be sized for it.
Shared SAN or clustered storage for larger facilities where many editors work from the same pool and the network can no longer carry it.
The common mistake is a NAS on 1GbE. It will hold the footage and it will not play it back to more than one editor at a time.
Which drives for production storage?
Two tiers, and the mistake is using one for both.
Active tier: SSD, either enterprise SATA/SAS SSDs in an array or NVMe where the platform supports it. This holds the working set for current projects and delivers the throughput. Our SSD versus HDD guide covers when flash pays, and production working sets are the clearest case.
Capacity tier: enterprise CMR hard drives in a redundant array. This holds completed projects, raw footage not currently in use, and archive. Our SMR vs CMR guide covers why CMR is non-negotiable here — large sequential writes are exactly where SMR looks fine until an array rebuild.
Mechanical drives in the active tier is the classic under-build. It works with one editor on compressed footage and fails the day a second editor opens a high-bitrate project.
RAID for video: which level?
Sustained throughput and rebuild time both matter.
RAID 6 or equivalent for the capacity tier, because large drives take a long time to rebuild and a second failure during rebuild loses everything. Our RAID guide covers the trade-offs.
RAID 10 or a flash array for the active tier, where write performance matters and rebuild is fast.
And the rule that applies everywhere: RAID is not backup. A production facility that loses a project to a double-drive failure with no backup has lost the project. Our backup guide covers the principle, and for production the backup target is frequently a second capacity array or tape.
Archive: where finished projects go
The tier that gets forgotten until the storage is full.
Finished projects need to leave the active and capacity tiers, or those tiers fill and performance degrades. Two options.
LTO tape, which is cheap per terabyte, lasts decades on the shelf, and is the industry standard for production archive. Our tape guide covers what a tape setup involves.
A separate archive array of high-capacity drives, which is faster to retrieve from but costs more to keep spinning.
Either way, the archive needs a catalogue. A project on tape that nobody can find is not archived.
The network matters as much as the storage
For any shared setup, three things.
10GbE minimum to each editing workstation, with 25GbE or better where high-bitrate formats are the norm. Our 10GbE adapter guide covers the workstation side.
A switch with enough backplane capacity to carry every editor at once, not just the port speed. Our switch guide covers what to check.
Cabling that supports it. 10GbE over old copper runs is unreliable. Our cabling guide covers what each category carries.
Frequently asked questions
What kind of storage is best for video editing?
Two tiers: SSD or NVMe for the active working set, and enterprise CMR hard drives in a RAID 6 array for capacity. Direct-attached for one workstation; a NAS on 10GbE or better for a small team; shared SAN for larger facilities.
Can I edit video from a NAS?
Yes, on 10GbE or faster with enough drives to serve the streams. A NAS on 1GbE will hold footage and will not play it back to more than one editor at a time. The network is usually the bottleneck before the drives.
Should I use SSD or HDD for video editing?
SSD for the active tier, HDD for capacity and archive. Mechanical drives in the active tier work with one editor on compressed footage and fail when a second editor opens a high-bitrate project.
What RAID level is best for video?
RAID 6 for the capacity tier, since large drives rebuild slowly. RAID 10 or a flash array for the active tier. And RAID is not backup; a double failure with no backup loses the project.
How should I archive finished video projects?
LTO tape for cheap, long-lasting shelf archive, or a separate high-capacity array for faster retrieval. Either way, keep a catalogue; a project nobody can find is not archived.
Can I use SMR drives for video storage?
Not in any array. Large sequential writes are exactly where SMR looks fine until a rebuild, at which point the array can fail. CMR only for production storage.
Tell us your footage format, editor count and project sizes, and we will size the two tiers and the network honestly.
