Storage is the quietest failure in a surveillance system. Cameras that stop working get noticed within days. Storage that is too small never announces itself — it simply overwrites, and the one recording you eventually need turns out to be four days older than the window you kept.
This guide shows how to size storage properly: the arithmetic, a reference table for bitrates, worked examples, and the reasons real-world capacity is always less than the number on the drive label.
The only formula you need
Video storage depends on bitrate, not on megapixels. A camera streaming at 4 Mbps consumes the same space whether it is 4 MP or 8 MP — the resolution only matters because it changes the bitrate. The arithmetic:
1 Mbps of continuous recording = 0.45 GB per hour = about 10.8 GB per day.
So: Storage (GB) = number of cameras × bitrate in Mbps × 10.8 × retention in days, then add roughly 10% headroom so the system is never writing into a full volume.
Typical bitrates as a planning reference
These are planning figures for H.265 with smart compression enabled, at 15–25 fps on a moderately busy scene. Your actual bitrate is whatever is configured on the camera — always confirm it in the live stream statistics before ordering drives.
| Camera | H.265 (typical) | H.264 (typical) | Per camera per day, H.265 |
|---|---|---|---|
| 2 MP / 1080p | 1.5 – 2 Mbps | 3 – 4 Mbps | ≈ 16 – 22 GB |
| 4 MP | 3 – 4 Mbps | 6 – 8 Mbps | ≈ 32 – 43 GB |
| 5 MP | 4 – 5 Mbps | 8 – 10 Mbps | ≈ 43 – 54 GB |
| 8 MP / 4K | 6 – 8 Mbps | 12 – 16 Mbps | ≈ 65 – 86 GB |
Two things follow immediately. First, moving from H.264 to H.265 roughly halves your storage bill for the same picture — if the recorder and cameras both support it, this is the cheapest capacity you will ever buy. Second, jumping from 4 MP to 8 MP roughly doubles storage, so the resolution decision is also a storage decision.
Worked examples, 30-day retention, continuous recording
| System | Bitrate each | Raw requirement | Plan for (with headroom) |
|---|---|---|---|
| 8 cameras, 2 MP | 2 Mbps | ≈ 5.2 TB | 6 TB |
| 16 cameras, 2 MP | 2 Mbps | ≈ 10.4 TB | 12 TB |
| 16 cameras, 4 MP | 4 Mbps | ≈ 20.7 TB | 24 TB |
| 32 cameras, 4 MP | 4 Mbps | ≈ 41.5 TB | 46 TB |
| 32 cameras, 8 MP | 8 Mbps | ≈ 83 TB | 92 TB |
| 64 cameras, 4 MP | 4 Mbps | ≈ 83 TB | 92 TB |
Halve the retention and you halve the number; a 16-camera 4 MP site at 15 days needs about 10 TB rather than 21 TB. This is exactly why retention should be decided as a policy before anything is ordered.
Four reasons the real number is higher than the formula
1. Labelled capacity is not usable capacity
A drive sold as 8 TB presents roughly 7.3 TiB to the recorder, because the manufacturer counts in powers of ten and the system counts in powers of two. Expect to lose about 7–9% before anything else.
2. RAID costs capacity by design
Redundancy is not free: RAID 5 gives up one drive’s worth of capacity, RAID 6 gives up two, RAID 1 gives up half. If the system must survive a drive failure without losing footage — and on any site holding evidence it should — budget for it in the capacity, not as an afterthought.
3. Drive bays are a hard limit
The recorder’s bay count, not your budget, sets the ceiling. A 4-bay unit with the largest drives it supports has a maximum, and passing it means a second recorder or a storage expansion unit. Check the ceiling before choosing the recorder, not after the second phase is approved.
4. Events and audio add up
AI event clips, snapshots, licence-plate records and audio all consume space alongside continuous video, and logs and databases grow over time. Leave room for them.
Event-based recording: real savings, real risk
Recording only on motion or on an AI event can cut storage dramatically — often by half or more on quiet sites such as warehouses at night. The trade-off is that anything the detection misses does not exist afterwards. The usual compromise is a hybrid: continuous recording on the small number of cameras that must never have a gap (entrances, cash handling, gates), event-based recording elsewhere, and a longer retention window on the continuous ones.
Why the class of drive matters more than the capacity
A surveillance recorder writes continuously, from every camera, forever. That is a fundamentally different duty cycle from a desktop PC, and it is why desktop drives fail early in recorders — usually within the first two years, and usually without warning.
What to look for when comparing drives:
- Rated annual workload (TB/year). Surveillance drives publish it and it is high; desktop drives are rated for a small fraction of it. Compare your system’s annual write volume against the rating — the formula above gives you the number.
- Designed for 24/7 write with many simultaneous streams, and firmware tuned for video rather than for file access.
- Supported camera/stream count for the drive model, and whether it is rated for AI recorders, which add analytics streams on top of video.
- Vibration tolerance in multi-bay enclosures — several drives in one chassis interfere with each other mechanically.
As the authorized Western Digital agent in Egypt, we supply WD Purple surveillance drives alongside Hikvision recorders, and we size the two together rather than selling capacity by itself. Equipment supplied by Speed4Trading carries a one-year warranty.
A sizing checklist you can send to any supplier
- What bitrate per camera was assumed, and with which codec?
- How many days of retention does that deliver — continuous or event-based?
- Is the figure usable capacity after formatting and after RAID?
- How many bays does the recorder have, and what is its maximum supported drive size?
- What is the drive’s rated annual workload, and how does it compare with this system’s annual write volume?
- What happens when the volume fills — overwrite, or stop?
Related reading
Storage is one of seven lines in a real quote — see what actually drives the cost of a CCTV system in Egypt. Because drives are the least inspected part of any order, read how to verify genuine equipment before you pay. And for what a modern recorder does with all that footage once it is stored, see AcuSeek NVR. At scale, Americana Group’s 450+ branches shows why retention policy has to be decided centrally.
Frequently asked questions
Does a higher-megapixel camera always need more storage?
In practice yes, because it produces a higher bitrate. But the bitrate is what you pay for, not the megapixels — a well-configured 4 MP camera with smart compression can consume less than a badly configured 2 MP one.
Is 30 days the right retention period?
Thirty days is a common default because incidents are often discovered weeks later, but it is a business decision rather than a technical one. The right question is: how late could someone realistically report something, and do we still want the footage then?
Can I add storage later?
Only up to the recorder’s bay count and maximum supported drive size. This is why the recorder should be chosen for the final system, not the first phase.
Do I need RAID?
If losing footage to a single drive failure is unacceptable, yes — and then budget the extra capacity it consumes. On sites where the video is convenience rather than evidence, it may not be worth it.
Can you check whether our current system’s retention matches what we were promised?
Yes. Send the camera count, resolution, codec and drive configuration and we will tell you the real retention window the system delivers.
Size it once, properly
Speed4Trading has been the authorized Hikvision and Western Digital agent in Egypt since 1999. Tell us your camera count, resolution and the retention you need, and we will calculate the storage — before anything is ordered.
Hotline 16454 · WhatsApp 01068835505 · Request a free consultation · Explore our security solutions