Retention of surveillance recordings in the company: how to determine it

Retention of monitoring recordings is not a technical parameter, but a decision that has both a legal and hardware layer. This article shows how to determine it based on the goal and not on the disk capacity, what determines the space consumption and what you need to remember before choosing a recorder.

Disclaimer. This article describes the technical and organisational side of retention selection. Not intended as legal advice. Determining the legal basis, purpose of processing and permissible storage period requires an assessment of a specific case by a lawyer or data protection officer. If monitoring covers employees, there are also separate obligations arising from labour law.

Retention of surveillance recordings is a decision, not a setting

A typical conversation about monitoring begins with the question of how many days of recordings will fit on the disks, although this is a question asked in the wrong direction.

The correct order is the reverse because it is first determined why recordings are stored and as long as they can remain in the system and only then selects the equipment that will implement it.

The reason is practical, not only formal, because when you first buy a recorder, you match the purpose to its capacity, i.e. you reverse the actual logic. The principle of storage limitation requires that personal data be kept no longer than is necessary for the purposes of processing, and a camera recording in which a person can be recognised is such data.

First the target

Before any number of days expire, it is worth answering four questions that organise both the organisational and technical layers:

What is this monitoring for: property protection, personal safety, process control and documentation of on-ramp events must be considered separately because different objectives justify different periods.

How much time actually passes before the event is noticed: is a key question that is often overlooked because a theft from a warehouse may only come to light during an inventory, while a parking incident is usually reported sooner. Retention shorter than the typical detection time means that monitoring fails to achieve its purpose.

Does monitoring cover employees: if so, a separate legal regime applies because Polish labour law regulates monitoring in the workplace separately, including the issue of storage period and information obligations towards employees and new hires. Check the current wording of the regulations and consult them with a lawyer, as this is not an area where it is worth relying on an industry article.

What happens to the recording secured for the purposes of the proceedings: material saved as evidence is subject to different rules than current recordings and should not be erased by automatic overwriting, so a procedure is needed to remove it from the normal cycle.

Only after obtaining these four answers does it make sense to talk about disks and system parameters.

What does space occupancy depend on?

The technical side is simpler, although it is sometimes mistakenly reduced only to image resolution.

Space occupancy depends on camera data stream multiplied by the number of devices and time, while the stream itself is due to several factors:

Factor	Impact
Resolution	more pixels, larger stream
Frame rate	usually linear
Compression codec	newer codecs produce a smaller stream with similar quality
Scene complexity and variability	movement, rain, shaking vegetation, changing lighting increase the flux
Recording mode	continuous, on detection, mixed

Most common surprise: A camera aimed at a busy stage generates a much larger stream than an identical device observing an empty hall with the same settings, because video compression takes advantage of the differences between frames, so a static image takes up less space than one full of motion.

This has two practical consequences, since an enumeration based on a catalogue stream remains an estimate, not a guarantee, and retention calculated in summer may not be accurate in autumn, when rain falls and swaying branches will increase the stream, causing previous recordings to be overwritten faster.

Recording mode changes the equation

Continuous recording provides predictable busyness because a similar amount of data is created each hour, which makes calculations easier but requires the most capacity.

Detection recording takes less time, but space usage becomes unpredictable as it depends on the number of events, and has a significant drawback: if the detection does not work, there is simply no recording. What is created is not empty material that can be viewed later, but a gap.

In practice, a mixed mode is often used, combining continuous recording at lower quality with higher quality at detection moments, thus creating a compromise between reliability and capacity.

The choice of mode is a decision about risk, not only about savings, so it is worth making it consciously and writing down the accepted justification.

How to calculate it

You can go through the following method yourself, keeping the order from target to infrastructure:

1. Set a goal and allowable period (section on purpose, with legal consultation).
2. List cameras and assign each scene a character: static, moderate, busy.
3. Set mode and parameters for each: resolution, frame rate, continuous or detection.
4. Take the stream from the data sheet cameras for these parameters.
5. Count daily occupancy for each camera and sum.
6. Multiply by the number of days retention.
7. Add stock. For a scene more variable than expected, for a season, for adding a camera, for material secured outside the cycle.

The seventh point determines whether the calculation will survive the impact with reality, because an installation calculated without reserve maintains the declared retention only until the first change in conditions.

It is also worth checking the opposite, i.e whether the recorder will be able to handle the total stream from all camerasbecause disk capacity and write performance are two different constraints.

Storage

Before selecting disks, it is worth knowing a few rules that affect safety and actual capacity:

• surveillance drives are a different category than office drives. The recording from the recorder is continuous, multi-stream and continues without interruption. Manufacturers produce series designed for such loads and it is worth using them,
• disk failure means loss of recordings. If material is important, consider a redundant configuration. This reduces the usable capacity, so it must be taken into account in the calculation,
• overwrite works automatically and does not ask for consent. The oldest recordings disappear when there is no space left. Therefore, the protected material must be removed from the cycle manually and consciously,
• actual retention is sometimes shorter than declared. It is worth checking after a month and after the change of season how many days are actually available.

The last point distinguishes a monitored installation from an abandoned one, because the recorder does not have to report that it has stored the material for much less time than expected.

What's worth writing down

Regardless of the technical side, it is worth writing down the findings so that they can be verified later:

• monitoring purpose and processing basis,
• established retention period and its justification,
• who has access to the recordings and to what extent,
• procedure for securing material for the purposes of the proceedings,
• procedure for responding to requests from persons to whom the recordings concern,
• surveillance scope: what cameras cover and what they intentionally don't.

The last point also has a technical dimension, because if the camera covers an area that it should not observe, you can limit the frame by masking the zone in the configuration. This solution may be easier than moving the device, so it is worth checking it upon receipt.

The selection of the frame and field of view is described in the article about camera placement and blind spots, separating image design from retention calculations.

Most common errors

• setting retention on disk capacity, not on target,
• retention shorter than the typical event detection time in a given company,
• no stock and no post-season verification,
• no procedure for securing the material, so the evidence disappears when overwritten,
• detection-only recording where a gap in the material is unacceptable,
• no arrangements have been made as to who has access to the recordings,
• treating employee monitoring as property monitoring.

How we do it

We start by talking about the goal and the time that elapses to detect an event in this particular company, and only then do we count the streams, select the capacity and mode, and show the accepted inventory.

We leave the legal layer to a lawyer or data protection inspector, preparing the technical description that such a person needs: scope of cameras, settings, method of overwriting and access rules. We do not replace this assessment and do not represent that the installation is compatible with anything.

After implementation, monitoring can be monitored continuously, including checking the actual retention, disk condition, recording efficiency and response before an event turns out that there are no recordings. This is the element that most often fails in installations left to their own devices.

Scope includes building and maintaining corporate networks, including IP monitoring installations combined with constant supervision.

Book a free IT review And let's talk about how many days of footage you really have.

Sources

• EDPB, Guideline 3/2019 on the processing of personal data by video devices: Polish version, adopted July 10, 2019, final version January 30, 2020
• Personal Data Protection Office: Polish version of the guidelines and national positions