CCTV Storage Solutions — NVR, SAN, NAS & Cloud

Storage is the silent foundation of every CCTV system. When it works correctly, nobody notices it. When it fails — footage is lost, retention periods are not met, drives crash, or recordings are overwritten before incidents are discovered — the consequences range from regulatory non-compliance to a complete inability to investigate security events. Getting storage right requires understanding both the technology options and the mathematics behind capacity planning.

1. Why Storage Is the Most Critical Infrastructure Decision

Consider the numbers: a single 4MP IP camera recording continuously at 15 frames per second using H.265 compression generates approximately 43 gigabytes of data every day. A 64-camera installation produces nearly 2.8 terabytes daily. Over a 60-day retention period, that translates to over 165 terabytes of raw video data that must be reliably written, stored, protected against drive failures, and made instantly searchable for playback.

Storage is also the single largest recurring cost component in a CCTV system. Hard drives have a finite lifespan (typically 3–5 years under continuous surveillance workloads) and must be proactively replaced. Under-provisioning storage means footage is overwritten before the mandated retention period expires. Over-provisioning wastes capital on unnecessary drives. A calculation-based approach — not vendor guesswork — is essential.

2. Storage Architecture Types

There are four primary storage architectures used in CCTV systems today. The right choice depends on the size of the installation, the required retention period, the budget, and the level of redundancy needed.

Storage Architecture Comparison

3. Storage Architecture in an Enterprise CCTV System

The following diagram illustrates how storage integrates into the overall CCTV network architecture, showing the data flow from cameras through the recording layer to tiered storage.

4. Storage Capacity Calculation

Accurate storage calculation prevents both under-provisioning (footage lost before retention period expires) and over-provisioning (wasted budget). The calculation is straightforward once you understand the variables.

The Five Variables

1. Number of cameras
2. Average bitrate per camera — determined by resolution, frame rate, compression codec, and scene complexity
3. Recording mode — continuous (24/7) recording uses the most storage; motion-triggered recording uses significantly less, but may miss critical events in the seconds before motion is detected
4. Retention period — the number of days footage must be kept before overwriting (typically 30 days for commercial, 60 days for PSU, 90+ days for banks)
5. RAID configuration — the redundancy scheme that protects against drive failures, but reduces usable capacity

Storage Calculation Formula

Bitrate Reference by Resolution and Codec

Worked Example — 64-Camera PSU Building

64 cameras, 4MP resolution, H.265 compression at 15 fps, continuous recording, 60-day retention, RAID 6:

• Average bitrate per camera: 4 Mbps (using the mid-range value for 4MP H.265)
• Daily storage per camera: 4 × 10.8 = 43.2 GB/day
• Total daily all cameras: 43.2 × 64 = 2,765 GB/day (2.77 TB/day)
• 60-day raw storage: 2.77 × 60 = 166.1 TB
• RAID 6 adjustment: 166.1 ÷ 0.67 = 247.9 TB
• File system overhead (+10%): 247.9 × 1.10 = ~273 TB total physical disk capacity
• In hard drives: approximately 17 × 16TB drives or 14 × 20TB drives (plus hot spares)

5. RAID Configurations for CCTV

RAID (Redundant Array of Independent Disks) combines multiple physical hard drives into a logical unit that provides either improved performance, data redundancy (protection against drive failure), or both. For CCTV, RAID is not optional — it is essential. Without RAID, a single drive failure means permanent loss of all recordings stored on that drive.

6. Surveillance-Grade Hard Drives

CCTV storage demands a specialised type of hard drive. Standard desktop drives — designed for 8 hours/day operation with intermittent read/write — will fail prematurely under the 24/7 continuous sequential write workload of video surveillance. Surveillance-grade drives are engineered specifically for this purpose.

Key Differences from Desktop Drives

Drive Replacement Planning

Surveillance drives have a typical operational life of 3–5 years. For a system with 20 drives, you should expect to replace 4–7 drives per year as they age. Budget for proactive drive replacement as part of the Annual Maintenance Contract (AMC). Monitor drive health using S.M.A.R.T. data — replace drives showing early warning signs (reallocated sectors, pending sectors, uncorrectable errors) before they fail completely.

7. Edge Storage — On-Camera Recording

Most modern IP cameras include a micro-SD card slot that enables the camera to record video locally on the camera itself, independent of the NVR or recording server. This feature — called edge storage — provides a critical safety net.

Automatic Network Replenishment (ANR)

ANR is the most important use of edge storage. When the network connection between the camera and the recording server is interrupted — due to a switch failure, cable damage, network congestion, or server downtime — the camera automatically begins recording to its local micro-SD card. When the network connection is restored, the camera automatically transfers the locally recorded footage to the central recording server, filling in the gap seamlessly. The operator sees a continuous, uninterrupted recording timeline.

Edge Storage Best Practices

• Use high-endurance micro-SD cards rated for continuous write operations (not standard consumer cards, which will fail within weeks). Industrial-grade or surveillance-rated cards from reputable manufacturers are essential.
• Card sizes of 128 GB to 256 GB are typical — providing 1–3 days of buffer recording depending on bitrate settings.
• Configure the camera to record the sub-stream (lower resolution) to the SD card rather than the main stream — this extends the available recording time significantly.
• High-endurance micro-SD cards have a typical lifespan of 1–3 years under continuous write workloads. Include card replacement in the AMC schedule.

8. Cloud & Hybrid Storage Models

Cloud-based video storage — where footage is stored on remote servers operated by a third-party provider rather than on on-premises hardware — offers compelling advantages for certain deployment scenarios but comes with significant constraints for Indian enterprise and government installations.

Full Cloud Storage

In a full cloud model, cameras stream video directly to cloud servers via the internet. There is no on-premises NVR or storage hardware. This model is attractive for retail chains, distributed small offices, and organisations without on-site IT infrastructure. Advantages include zero hardware maintenance, automatic scaling, built-in redundancy, and access from anywhere. Disadvantages include ongoing subscription costs (which can exceed on-premises costs over 3–5 years), dependency on internet connectivity (footage is lost during outages), significant upload bandwidth requirements, and potential data sovereignty concerns.

Hybrid Storage (Recommended for Enterprise)

The hybrid model — local recording with selective cloud backup — is emerging as the optimal approach for Indian enterprise deployments. Video is recorded locally on-premises (NVR, server, or NAS) for full-quality retention, ensuring recording continues even during internet outages. Simultaneously, selected content is uploaded to the cloud: alert clips triggered by analytics events, low-resolution sub-streams for remote viewing, or periodic backup copies of critical cameras. This provides the reliability of local recording with the accessibility and disaster recovery of cloud.

Data Sovereignty for Indian Organisations

For PSU and government CCTV installations, any cloud storage component must comply with Indian data localisation requirements. The Digital Personal Data Protection Act, 2023 (DPDPA) has implications for cross-border transfer of personal data — and CCTV footage of identifiable individuals qualifies as personal data. Verify that the cloud provider stores data exclusively within Indian data centres and complies with applicable Indian regulations.

9. Storage Sizing Quick Reference by Building Type

10. Storage Maintenance & Lifecycle Management

Ongoing Monitoring

• Daily: Verify all cameras are recording. Check storage utilisation dashboard — ensure retention targets are being met.
• Weekly: Review drive health reports (S.M.A.R.T. data). Check for any degraded RAID arrays.
• Monthly: Verify that the oldest recordings match the expected retention period. Test footage playback and export from the oldest available date.
• Quarterly: Review storage growth trends. Plan capacity expansion if camera additions are expected.

Proactive Drive Replacement

Do not wait for drives to fail before replacing them. Surveillance drives have a rated operational life of 3–5 years. After year 3, begin proactive replacement of the oldest drives in the array. This scheduled replacement — during planned maintenance windows — avoids the risk of emergency drive failures during critical recording periods. Always maintain at least one hot spare per RAID array, and keep two additional spare drives in inventory for immediate replacement.