Retention planning
A 32TB drive does not have one universal “number of days.” Retention is controlled by camera count, actual bitrate, codec, frame rate, scene complexity, recording schedule and the space reserved for the recorder. Use the calculator with measured or configured bitrate values.
Quick answer
For reliable planning, use the average bitrate reported by the cameras or VMS rather than a resolution label such as “4K.” Two 4K cameras can consume very different storage if their codecs, frame rates and scenes differ.
Exact models in this guide
Seagate SkyHawk AI
ST32000VE000AI-enabled NVR and edge video analytics
Interactive planner
32TB recording-time calculator
Why bitrate controls retention
Storage grows from bits written over time. Camera resolution influences bitrate, but it does not define it by itself. A quiet hallway with variable bitrate can consume far less than a busy road at the same resolution.
Use Mbps from the camera or recorder settings whenever possible.
Retention planning improves when you export actual bitrate statistics from the recorder after the cameras have operated through representative busy and quiet periods. Use that measured average to replace the initial design estimate. The calculator then becomes a living capacity model rather than a one-time guess based only on resolution and frame rate labels.
Camera count multiplies storage
Eight cameras at 8Mbps behave like a combined 64Mbps stream before secondary streams, metadata or event clips. Doubling camera count roughly halves retention if every other setting stays the same.
The calculator multiplies the average bitrate by the number of continuously recorded streams.
When cameras use different bitrates, calculate a weighted total instead of multiplying one average by every camera. The simplified calculator assumes one average value for speed, but a detailed design can sum each configured stream. This matters when a few high-detail 4K cameras dominate storage while many lower-resolution cameras contribute relatively little.
Continuous versus motion recording
A motion-only system does not write at full bitrate 24 hours per day. The duty-cycle field lets you estimate the percentage of time cameras are actively recording.
Use conservative assumptions. A retail entrance or road can have far more motion than an office corridor.
Motion duty cycle should come from the site, not a universal percentage. A warehouse aisle at night may record rarely, while a city entrance can trigger almost continuously. If motion-only recording is mission critical, measure the actual event ratio after commissioning and keep enough reserve for seasonal changes in traffic.
H.264 versus H.265
H.265 can reduce bitrate for similar visual quality when the camera and recorder implement it well, but the saving is not a fixed percentage for every scene.
Instead of applying a generic codec discount, enter the actual configured or observed bitrate after choosing the codec.
Codec changes should be validated visually as well as numerically. Reducing bitrate aggressively can extend retention while creating compression artifacts that make identification harder. Storage planning should therefore start after acceptable image quality is established, not use storage savings as the only objective.
Reserve recorder headroom
Do not allocate the full marketed capacity to video. Recorders also need databases, indexes, thumbnails, metadata, logs and working space for overwrite behavior.
The reserve field reduces the capacity available to recordings before retention is calculated.
Recorder headroom also protects normal overwrite behavior. Systems that operate permanently near full capacity can spend more time managing indexes and deletions. Follow the vendor’s recommended free-space policy when one exists, and use the reserve field to model that requirement instead of counting every byte toward retained footage.
32TB is decimal capacity
Manufacturers state capacity in decimal terabytes. Operating systems often display binary tebibytes, so the visible number is lower even before filesystem overhead.
That display difference is normal and does not mean storage is missing.
The decimal-versus-binary difference should be handled consistently when comparing capacities. The safest approach is to calculate from bytes or use the same convention for every candidate. Mixing decimal TB on one side and binary TiB on the other can make a retention comparison appear more precise than it really is.
AI metadata also consumes space
Object detections, thumbnails, event clips and secondary streams can add storage beyond the main video stream. The amount depends on the VMS architecture.
If metadata storage is material, reserve additional capacity rather than pretending the video bitrate is the entire workload.
Analytics metadata can be modest in one VMS and substantial in another, especially when the system saves snapshots, face crops, license-plate images or duplicate event clips. Review the platform storage layout. If metadata is stored on the same volume, include a reserve or a measured daily overhead in the design.
Use measured peak and average values
Average bitrate is best for retention estimates, while peak bitrate matters for network and write-performance planning. A VBR camera may briefly exceed its average by a wide margin.
Review recorder statistics after several representative days and update the estimate.
Peak bitrate matters when many cameras become busy at the same time. Even if average retention fits comfortably, the recorder, network and disk subsystem must absorb those bursts. Separate the capacity question from the throughput question so a long-retention design does not mask a performance bottleneck.
Retention policies need margin
If regulation or business policy requires 30 days, do not design for exactly 30.0 days under perfect conditions. Firmware changes, busier scenes and added cameras can erase that margin.
Target a buffer above the required retention period.
Policy margin should cover growth as well as uncertainty. If a site may add four cameras next year, model the future configuration now. Buying a drive that barely meets today’s 30-day target can force an early replacement even though the original disk remains healthy.
Multiple drives and RAID
If the recorder uses several 32TB drives, total retention depends on how the storage pool is configured. RAID parity or mirroring reduces usable capacity.
Use the RAID guide for multi-drive protected capacity rather than multiplying raw capacity blindly.
Multi-drive systems need a two-step calculation: determine protected usable capacity from the RAID layout, then apply the recording formula. Do not multiply 32TB by drive count and call the result retained storage when parity, mirroring or hot spares consume part of the raw total.
When 32TB is excessive
A small system with few cameras, low bitrates and short retention may not benefit from 32TB. A smaller SkyHawk AI capacity can cost less while comfortably meeting policy.
Choose capacity from measured demand rather than buying the largest drive by default.
A 32TB disk can be excessive in a small installation because unused retention does not improve video quality. The right-size decision can free budget for better cameras, redundancy, backup or UPS protection. Capacity should support the surveillance objective rather than become the objective itself.
Deployment check
Confirm ST32000VE000 compatibility, configure camera bitrate, estimate duty cycle, leave headroom and review actual retention after installation.
The calculator is a planning model, not a guarantee of exact days in every scene.
After deployment, compare the calculator estimate with the recorder’s observed “oldest footage” date. If actual retention is shorter, investigate real bitrate, event clips, secondary streams, reserve settings and added cameras. That feedback loop is more reliable than repeatedly adjusting capacity from theoretical assumptions.
Current exact-model offers
Live buying options for the models above
Only listings containing the exact model number can appear. Generic capacity-only matches, accessories, systems and mixed-model listings are excluded.
Current manufacturer references
Exact models, warranty terms and product specifications can change. These links are included so you can verify the current manufacturer record before purchase or deployment.
Questions people ask
Frequently asked questions
How many days can 32TB record?
It depends on total average bitrate, camera count, recording schedule and reserved capacity.
How much does one 8Mbps camera use per day?
Roughly 86.4GB/day before filesystem and recorder overhead.
Does 4K always use more storage than 1080p?
Usually, but bitrate settings and codec efficiency determine the actual difference.
Does H.265 always cut storage in half?
No. Savings vary by encoder, scene and quality settings.
Should I use peak or average bitrate?
Use average bitrate for retention and peak bitrate for throughput planning.
Does motion recording extend retention?
Yes when cameras record for only part of the day, but the duty cycle must be estimated realistically.
Does AI metadata use storage?
Yes, depending on the recorder and analytics architecture.
Why does 32TB show as about 29TiB?
Manufacturers use decimal TB while many operating systems display binary TiB.
Does RAID reduce recording days?
Yes when capacity is consumed by mirroring or parity.
Should I plan exactly to the required retention days?
No. Leave a safety margin for bitrate changes, added cameras and system overhead.