Home / Engines / Drive Endurance Fit

Will this drive survive your camera wall?

Describe the recording — cameras, resolution, frame rate, codec, retention — and either name a drive or let the registry rank every model that matches your interface and media. The engine turns the camera bitrate into TB/day and TB/year, holds it against the endurance the vendor publishes (an HDD's annualised workload rating, an SSD's terabytes written or drive writes per day), checks the retention footprint against the capacity, the service life against the endurance budget and the peak write rate against the datasheet, and returns FIT, FIT WITH RISKS, NEEDS VALIDATION or NO FIT — with the model that would fit instead.

01 · Describe the recording

Cameras & streams
Cameras
4
8
16
32
64
Custom count
Resolution
720p
1080p
4K
Frame rate
10 fps
15 fps
25 fps
30 fps
Codec
H.264
H.265
Recording & retention
Recording
Continuous
Motion triggered
Retention
7 days
14 days
30 days
60 days
90 days
Custom retention
Drive
Media
Interface
Model
Loading drive registry…
Advanced
Drives per node
1
2
4
8
Service life
3 years
5 years
7 years
10 years
Motion share
0.1
0.3 (default)
0.5
Write amplification
1.1 (default)
1.3
1.5
Ready — ⌘↵ to run

03 · How this works

Four checks on one write rate

The camera wall's average bitrate comes from the same table the Network Bandwidth and Storage Endurance engines use. Multiplied by the recording duty cycle and the write amplification it becomes TB/day and TB/year; the un-amplified rate times the retention window is the capacity the recording needs. Capacity holds that footprint against the drives in the node. Endurance holds the annual load per drive against the vendor's figure — an HDD's annualised workload rating in TB/year, an SSD's terabytes written spread over its warranty, or drive writes per day when no TBW is published. Service life divides the endurance budget by the annual load and compares the result with the years you plan to run the node. Write throughput holds the peak write rate against the datasheet's sequential-write or sustained-transfer figure.

Categorical, and honest about gaps

The verdict is FIT, FIT WITH RISKS, NEEDS VALIDATION or NO FIT — never a predicted failure date. Every drive figure is class A (the vendor datasheet, quoted and linked); the write amplification (1.1) and the motion share (0.3) are class E engineering defaults and are labelled as such in every result. A figure the datasheet does not publish is UNKNOWN, which is why some rows come back NEEDS VALIDATION rather than a number.

Full rules, the class E defaults and what invalidates a result: methodology. Raw dataset: datasets/drive-endurance.json. Related: Storage Endurance sizes the array; Network Bandwidth sizes the link that feeds it.

04 · FAQ

Does this predict when the drive will fail?
No. It compares the write load your cameras produce against the endurance the vendor publishes — an HDD's annualised workload rating in TB/year, or an SSD's terabytes-written budget spread over its warranty. That tells you whether the drive is inside or outside its specification, and how many years its endurance budget covers at this load. Field failure also depends on temperature, vibration, power events and firmware, none of which a datasheet figure captures.
Why is write amplification 1.1 and not something else?
1.1 is an engineering default (class E) for the filesystem and container overhead on a continuous video write path — large sequential writes with little metadata churn. It is applied to the endurance side only; the retention footprint uses the un-amplified rate, because the amplification is written and discarded rather than retained. Override it in Advanced if you have measured your own.
Where do the drive numbers come from?
Vendor datasheets and product briefs: Western Digital Purple and Purple Pro, Seagate SkyHawk, SkyHawk AI and IronWolf Pro, Samsung 990 PRO and 870 EVO, Kingston NV3, DC2000B and DC1500M. Every row names its document with a verbatim quote and a link. A figure the datasheet does not state is left null and the check reports UNKNOWN rather than guessing.
What is the difference between the endurance check and the service-life check?
The endurance check is a rate: TB/year written per drive against the TB/year the vendor rates. The service-life check is a budget: the total endurance the drive carries divided by the annual load, against the years you plan to run it. An HDD's workload rating is a rate ceiling that has to hold every year; an SSD's TBW is a budget that has to outlast the deployment.