What is edge recording and how does it work?
Edge recording writes video to storage located at the site, such as a drive in a recorder or a card in the camera. The encoder compresses frames, and the system writes them continuously or only when triggered, overwriting the oldest footage once storage is full. This is called loop recording. Its strength is independence from the network. If the cellular link slows or drops, recording continues and nothing is lost, so the full-quality footage is available later.
The weaknesses are physical: if the equipment is stolen, burned or damaged, the footage goes with it, and reviewing footage means retrieving it remotely, which can be slow over limited links. Storage media wear out. Drives and cards have limited write endurance, so continuous recording suits industrial-grade media better than consumer parts, and checking health periodically avoids discovering failure after an incident. The simplest way to think about it is that edge recording favors completeness and independence, while cloud recording favors safety and access.
What is cloud recording and what does it need?
Cloud recording uploads video to servers in remote data centers, where it is stored, indexed and made available through a web interface. The footage is outside the physical site, so it survives anything that happens to the equipment. Platforms can also run analytics and share clips without anyone visiting the site. The cost is bandwidth and dependence. Continuous full-quality upload demands a lot of sustained uplink, which at cellular sites is the scarce resource.
If the link goes down, uploads stall or are lost unless something buffers them. Platforms therefore usually upload reduced streams, event clips or short continuous windows instead of everything, in full quality, all the time. Security of the cloud account is also part of the picture, including who holds credentials, how access is logged and whether footage is encrypted while moving and while stored. Cloud services also make it easier to share clips with colleagues or authorities, and to keep a searchable history without visiting the site.
Why use both in a hybrid design?
A hybrid design records full-quality video locally and sends selected material to the cloud. Local storage provides the complete record and a safety net during outages. The cloud copy of important moments protects the evidence if the site is tampered with, and supports remote verification of alerts. The two layers complement each other's failure modes. Network problems do not stop the local recording, and site damage does not erase what was already uploaded.
When evaluating a hybrid, ask what is uploaded by default, how long each layer retains footage, whether footage that missed upload during an outage is sent later, and how you retrieve full-quality clips from the site on demand. Smart buffering can bridge gaps: a unit that records locally and forwards missed clips when the link returns turns an outage from a loss into a delay, though the forwarding itself consumes bandwidth. The point of combining them is not duplication for its own sake, but making sure that a single failure never costs you the footage you need.
How do you size storage and retention?
Storage is arithmetic. A stream's bitrate, the data per second, multiplied by the seconds you want to keep, gives the bytes required for one camera. Multiply by the number of cameras. Resolution, frame rate, compression and scene activity all change the bitrate, and modern compression such as H.265 reduces it for similar quality. Recording only on motion or analytics events cuts storage but risks missing context before the trigger, so systems keep a short pre-event buffer.
Retention is a policy choice: longer retention helps investigations that surface late, but costs more storage. Check with counsel or your records policy about how long footage must or must not be kept. Retrieval speed is the practical test. Footage that exists but takes hours to reach the person investigating is far less useful than a short clip delivered within minutes of an alert. Storage planning should also include the overhead of audio, metadata and pre-event buffering, which add to the raw video requirement.
What should you ask a provider about recording?
Ask what happens when the connection drops, and where the footage lives at each stage. Ask who can access it, how it is protected in transit and at rest, and how evidence is exported with its time stamps intact. Chain of custody matters if footage may be used by police or in court. Ask also about limits. Some services charge by data volume, throttle after a threshold or cap retention. Understand whether pricing models encourage you to record less than you need.
The best arrangement is one you can explain simply: what is stored locally, what goes to the cloud, how long each is kept and how to retrieve it. Finally, ask how footage can be exported and in what format, whether time stamps and camera names are preserved, and whether you can hand it directly to the authorities who ask for it. A provider who can describe recording in a few plain sentences usually understands it; one who cannot is a signal to ask more questions.
Key terms
- Edge recording
- Storing video on equipment at the camera site, which keeps recording whether or not the internet connection is working.
- Cloud recording
- Storing video on remote servers reached over the internet, protecting footage from site damage and enabling remote access.
- Loop recording
- Continuous recording in which the oldest footage is overwritten automatically once available storage is full.
- Hybrid recording
- Combining local full-quality recording with selected cloud uploads to gain resilience and remote access together.
- Pre-event buffer
- A short rolling memory of video kept so a triggered clip includes moments before the trigger occurred.
- Retention
- How long recorded video is kept before deletion, set by storage capacity, policy and legal requirements.
A trailer like this works where connectivity is imperfect, so both layers help. Standard value-adds include edge recording on the unit and 2 weeks of cloud recording, with unlimited data and no overages. AI VisionStream, VDS's cloud video management platform, holds the cloud copy, while the unit keeps recording if the cellular link drops. These details are part of what a buyer should ask about.
Common questions
Is cloud recording better than edge recording?
Neither is simply better. Cloud protects against site loss and eases access; edge protects against network loss and carries full quality. Using both usually gives the most dependable result. Consider the cost of storage over time, including the effort of retrieving footage, not only the price of the equipment.
What happens to footage during an internet outage?
With edge recording, the unit keeps saving video locally. Whether it later uploads missed material depends on the platform, so ask how gaps are handled. Edge storage also supports review of events that were never flagged, because the full record is still there.
How long should I keep security footage?
It depends on your risk, storage and obligations. Confirm any legal or policy requirements with counsel or the local authority, and keep footage tied to incidents longer than routine recordings. Both layers should be periodically tested by retrieving a real clip, since an untested archive is only a hope.