Somewhere in the sales process for every commercial camera system, this question shows up: should your footage live on a recorder in your building, or on someone else's servers?
Installers rarely give you a straight answer, because most only sell one of the two. We design and support both, so we don't have a horse in this race — and the honest answer is that neither model wins outright. Each is the right call for a specific kind of site, and plenty of businesses are best served by a mix of the two.
What the difference actually is
Strip away the marketing and the two models differ on exactly two things: where the footage lives and who maintains the recorder.
On-premise recording means a network video recorder (NVR) — a purpose-built computer with hard drives — sits in your building, usually in a network closet or locked cabinet. Cameras stream to it over your own network, and the footage never has to leave the property. You (or your installer, under a service arrangement) own the hardware, replace drives as they age, and apply firmware updates.
Cloud recording means your cameras stream footage over your internet connection to the provider's servers. There's no recorder in your building, or only a small gateway. The provider owns the storage, patches the software, and gives you an app to review footage from anywhere, for a per-camera monthly fee.
Everything else flows from those two facts. So let's take it one dimension at a time — with a verdict by scenario rather than a winner, because the winner depends on who's asking.
1. Cost shape: capital up front vs subscription forever
We're deliberately not quoting dollar figures — pricing moves, and any number printed today misleads you next year. What doesn't change is the shape of the two cost curves.
An on-premise recorder is mostly a capital cost: hardware and installation once, then years of modest running costs. The longer you keep it, the cheaper each month of recording effectively becomes; the trade is a bigger first invoice and ownership of aging hardware. Cloud recording is a pure operating cost: little hardware up front, then a per-camera fee that never ends. Month sixty costs the same as month one — multiplied by every camera you've added since, and scaled up again by how many days of retention you keep.
Verdict by scenario: Own your building and plan to stay — the on-premise curve rewards you more the longer you hold it. Short lease, protecting cash flow, or preferring a predictable operating expense — the cloud curve is doing exactly what you hired it to do. The mistake is comparing month one of cloud against year one of on-premise; run the comparison over the years you expect to operate the system.
2. Internet dependency: what happens when the connection drops
An on-premise recorder doesn't care whether your internet is up. Cameras stream to it over your local network, so recording continues through a service outage — you lose remote viewing, not footage. Cloud recording depends on your upload connection twice over: for capacity (every camera pushes video upstream all day, and commercial plans usually have far less upload than download) and for availability — when the connection drops, footage can't reach the cloud.
Reputable cloud systems mitigate this with edge buffering: the camera or an on-site gateway holds recent footage locally and uploads the backlog when the connection returns. It's a real solution, but read the fine print on two numbers — how long the buffer holds, and what happens when it fills. A buffer that covers a two-hour outage does nothing for the three-day one after a storm, and BC's storm seasons produce those.
Verdict by scenario: Urban site with business-grade fibre and a cellular backup — cloud's internet dependency is a manageable, known risk. Rural property, or anywhere that loses power and internet together in winter — the on-premise recorder's indifference to your connection is worth a lot. Going cloud anyway? Get the edge-buffer duration in writing.
3. Retention flexibility and the storage math
Retention — how many days of footage you keep — is a decision most businesses make by accident, by accepting a default. Decide it on purpose, because the two models handle it very differently.
On an NVR, retention is arithmetic: drive capacity divided by the data your cameras produce per day — driven by camera count, resolution, frame rate, and continuous-vs-motion recording. Retention is effectively free once the drives are bought, but it's capped: when the drives fill, the oldest footage is overwritten, and every camera competes for the same pool. Add four cameras and every camera's retention shrinks unless you add storage.
In the cloud, retention is a plan tier. Want more days? Pay more per month — flexibility with a compounding price, because long retention is many days, times many cameras, times every month you operate. The upside: you extend retention with a billing change instead of a hardware visit.
Before choosing either, answer one question honestly: how long would it take you to discover a problem? Theft, shrinkage, and disputed incidents often surface weeks after the fact, and some insurers need weeks of history, not days.
Verdict by scenario: Long retention on a stable camera count — on-premise arithmetic favours you strongly. Retention needs likely to change — cloud's flexibility earns its fee. Either way, decide the days first, buy second.
4. Multi-site management
If you operate one building, this section barely applies. If you operate five, it might decide the whole question.
With cloud recording, every site's cameras land in one dashboard: one login, one permission system, one place to confirm every camera at every location is online and recording. Adding a site is an enrolment exercise.
On-premise recorders can absolutely be managed across sites — the tools exist and we deploy them — but it takes deliberate design and someone who owns the discipline of checking each system's health. Done casually, you get the classic multi-site failure: the recorder at the branch nobody visits stopped recording eight months ago, and you find out the day you need the footage.
Verdict by scenario: Single site — don't let this dimension drive the decision. Five-plus sites without a dedicated IT resource — centralized management is the strongest single argument in cloud's favour, and often the reason a hybrid design puts the management layer in the cloud even when bulk footage stays local.
5. Maintenance and updates: who patches the NVR
Recorders and cameras run software; software has vulnerabilities; someone has to apply the patches. An unpatched recorder is one of the most commonly exploited devices on a business network, so the question is who owns that job.
With cloud recording, the provider patches the platform continuously, with no site visit — and storage capacity, failing hardware, and upgrades are their problem too. Note the boundary: the cameras themselves still sit on your network and still need firmware updates. Cloud shrinks your maintenance surface; it doesn't eliminate it.
With an on-premise recorder, patching is your responsibility, or your installer's under a service arrangement. This is where systems quietly rot: the installer hands over an app, and nobody touches the firmware for five years. Honestly framed, this dimension isn't cloud vs on-premise — it's maintained vs unmaintained. A recorder under an active service plan is a well-kept system; the same recorder with nobody assigned to it is a liability with a lens.
Verdict by scenario: No IT staff and no appetite for a service relationship — cloud's hands-off maintenance is worth real money. Existing IT capability, or a support arrangement with your installer — on-premise maintenance is a solved problem. What you should never choose is a recorder with nobody assigned to keep it alive.
6. Data residency and privacy
With an on-premise recorder, "where is our footage stored?" has a one-word answer: here. It sits on drives in your building, subject to Canadian law, and goes nowhere unless you send it.
With cloud recording, footage lives in the provider's data centres — and where those are varies. Some providers store Canadian customers' video in Canada; others store it in the United States or across several countries. Data stored abroad is generally subject to that country's laws, including its lawful-access regimes — and BC businesses operate under Canadian and provincial privacy legislation governing how personal information is handled, which surveillance footage of identifiable people is.
We won't turn this into legal advice — residency obligations are a conversation for your counsel or privacy officer. As a practical matter: ask any cloud provider, in writing, where Canadian customers' footage is stored, whether Canadian-residency storage is available, and who inside their organization can access it. A provider with good answers gives them quickly; one that gets vague is telling you something too.
Verdict by scenario: Regulated activity or clients with residency requirements — on-premise is the simplest defensible posture, and cloud is viable only with verified Canadian residency in the contract. A typical retail or industrial site — a due-diligence question, not a disqualifier, but ask before you sign.
7. Failure modes: drive death vs subscription lapse
Every recording system fails eventually. The models fail differently.
On-premise systems fail physically. Hard drives are consumable parts — writing video around the clock wears them out, and a dead drive in an unmonitored recorder means the system sits there looking normal while recording nothing. The blunter failure: the recorder lives on the property, so a fire, a flood, or a thief who takes the recorder on the way out destroys the evidence along with everything else. The incident that erases its own footage is the strongest argument for putting some footage off-site.
Cloud systems fail administratively. No drive to die, but a subscription to lapse — an expired payment card, an invoice bouncing to a former employee's inbox — and recording stops just as silently. The dependency chain is longer too: camera, network, internet service, and provider platform all have to be healthy for footage to land.
The common thread is the word silently. Both failures are survivable if noticed within days and catastrophic if unnoticed for months. Whichever model you choose, the real specification is: who gets alerted when recording stops, and how fast?
Verdict by scenario: If your threat model includes the recorder itself being taken or destroyed, you need an off-site copy of your critical cameras — full stop. If your weakness is administrative follow-through, engineer around the subscription lapse. In both cases, buy the alerting, not just the storage.
The hybrid answer: why it's often not either/or
The two models fail in opposite directions — which is exactly why the strongest design for many commercial sites uses both.
The pattern we deploy most often: record everything on-premise, push what's critical to the cloud. An NVR captures every camera at full resolution with generous retention, no per-camera fees, and no dependence on your internet connection. The handful of cameras that would matter most in a serious incident — entrances, the till, the safe, the loading door — also stream to cloud storage, so an off-site copy survives anything that happens to the building, and the cloud fee stays small because it covers a few cameras, not all of them. Multi-site operators often add cloud-based management — one dashboard, one health view — while the heavy footage stays on local recorders.
Hybrid isn't automatically right either; a two-camera site doesn't need architecture, it needs a decision. But when a business can't choose between the models, it's usually because they genuinely need a property from each. That's not indecision — that's a hybrid spec.
A decision framework by business profile
Single-site retail or office. Start from an on-premise recorder — the economics favour it and your retention arithmetic is stable. Add cloud replication for your two or three critical cameras if a break-in could take the recorder itself (for street-facing retail, assume it could). Make sure someone owns drive health and firmware.
Multi-site operations. Start from the management problem, not the storage problem. Five-plus locations with no dedicated IT resource points to cloud-first or a cloud-managed hybrid — it's what prevents the branch-recorder-died-quietly failure. Verify data residency in writing, and price the subscription across your full camera count and planning horizon before signing.
Construction and temporary sites. No network closet, questionable power, a temporary internet connection — and the site moves. Cloud-connected cameras with substantial edge buffering fit this reality far better than a recorder in a trailer, and a month-to-month subscription matches a project's finite life. Ask hard questions about buffer duration.
Whatever your profile, settle these before anyone quotes hardware: how many days of retention you need, what happens when the internet drops, who maintains the recorder for the next five years, and who gets the alert when recording stops. A designer who asks those questions is designing a system; one who doesn't is selling you a box — or a subscription.
Frequently asked questions
Can we switch from one model to the other later?
Usually, but the ease depends on install-time decisions. Cameras built on open standards can feed either an on-site recorder or a cloud platform, and the cabling and mounting — most of the installation labour — carry over either way. Cameras locked to one provider's ecosystem can't migrate; switching means replacing hardware. If you might change models later, say so during design — it costs little to keep the door open and a lot to reopen it.
Does cloud recording mean other people can watch our cameras?
It means your footage sits on infrastructure the provider operates, protected by their security controls — not the same as being watchable by strangers, but not a drive in your locked closet either. Ask any provider: is footage encrypted in transit and in storage, which staff can access customer video, and is that access logged? Serious providers answer these in published security documentation. Your own hygiene matters just as much — a strong, unique password and two-factor authentication on the viewing account protect either model.
With an on-premise recorder, can we still view cameras remotely?
Yes — remote viewing and cloud recording are separate things, and conflating them is the most common confusion in this decision. An on-premise system can be set up for secure viewing from your phone anywhere; the footage stays on your recorder, and only the stream you're watching leaves the building. What you don't get without a cloud component is off-site storage. Don't choose cloud recording to get an app; choose it for where the footage lives.
Where the footage lives is your call — the design is ours
We design, install, and support both models, plus the hybrids that combine them — so our recommendation comes from your site, your retention needs, and your appetite for maintenance, not from which product we happen to sell.
If you're weighing this decision for your own building, book a systems consult and we'll walk your site, run the retention math with you, and give you a straight answer about which side — or which mix — your business actually needs.