Networks Are Like Life: Not Every “Gigabit” Is a Blessing

Marketing sold us a very simple fairy tale a long time ago: the bigger the number on the box, the closer you are to a bright digital future.

100 Mbps is for the poor.

1 Gbps is for people who “know tech.”

And then this very knowledgeable person installs a gigabit router at home or on a site, pulls a “gigabit-rated” cable, hangs IP cameras on all of it and genuinely wonders why, instead of cyberpunk perfection, the archive shows jittery video and cameras that vanish like shy ghosts. Surprise: in video surveillance, gigabit is not always an upgrade. Sometimes it’s just a great way to give your network a nervous twitch.

There’s also a widespread feeling that cameras need gigabit. In reality, that’s almost always an illusion. A typical IP camera with sane settings eats about 4–12 Mbps on H.264 and 2–8 Mbps on H.265. Even 4K models rarely go beyond 25–30 Mbps. For them, 100 Mbps is like an empty three-lane highway at 2 a.m. Sure, you can shut down the whole city for Formula 1, but the car won’t magically go faster. What will happen is that the network becomes more complex and more temperamental. That’s why in most reasonable projects, cameras happily live on 100 Mbps, while gigabit is saved for where it’s actually needed — backbone links and the system core.

The first problem is that gigabit is brutally honest about what your cable really is.

For 100 Mbps, two living wire pairs and a “close enough” crimp are sufficient. It’s the kind uncle who politely ignores twisted pairs in junction boxes, market-bought “Cat.5e” of mysterious origin, and the classic “I crimped it from memory, but it should be fine.”

Gigabit, on the other hand, is a pedant. It wants all four pairs, proper twists, correct geometry, and connectors wired by the book — not “how we usually do it.” On 100 Mbps everything seems fine: stable link, camera recording. Switch the port to 1 Gbps — and the fun begins. The link flaps, packets disappear, the image freezes from time to time. The user writes to support: “your software is buggy.” In reality, it’s the twisted pair that’s been used for ten years as an experimental testing ground.

The second trap is hiding inside the box labeled “Gigabit Router.” On the outside, it shines with marketing promises. Inside, there’s usually one overworked chip that has to do everything.

Inside a camera lives a SoC — a System-on-a-Chip — combining CPU, codecs, networking, and more into a single piece of silicon. In a router, it’s the same philosophy: one SoC that does NAT, Wi-Fi, IPTV, and suddenly is expected to handle another dozen or two constant video streams.

When it’s just “family internet,” it copes. When 10–20 camera streams start flowing through it 24/7 — no pauses, no ads, no mercy — the chip heats up, starts thinking deeply about life, and then simply begins dropping connections. In the camera interface everything looks “connected,” the Gigabit LED is proudly lit, but the NVR sees half the cameras… or none at all. Formally, you have a “gigabit network.” In reality, you’ve conscripted a home router into industrial labor. A gigabit label doesn’t magically turn consumer hardware into professional equipment.

A special genre is PoE plus gigabit. Budget PoE switches love living on the edge of their power supply and common sense. Add long cable runs, cold weather, voltage drops, and a gigabit PHY that consumes more power, runs hotter, and is more sensitive to cable quality and you get the nightly classic:

“Why do two outdoor cameras drop exactly at 3 a.m.?”

At 100 Mbps they behave like tanks: stable link, solid PoE, archive recording. Switch to 1 Gbps “to be modern” — and here we go: a camera reboots, disappears, comes back, but no longer answers ping. The funniest part? PoE power can be perfectly fine — everything is green in the switch UI — but the video stream still collapses from time to time. Electricity flows. Packets… not always.

Now, where everything converges into one point, gigabit stops being a luxury and becomes a necessity. A simple example: 24 cameras at 8 Mbps each is about 192 Mbps of pure video traffic, plus protocol overhead and some margin. Trying to live with a 100-Mbps uplink between the camera switch and the server is like leaving a shopping mall through a single gate on Black Friday. This is where gigabit shines: it pulls all streams together without creating a bottleneck and lets the system scale.

There’s also a special class of “monsters”: multi-sensor cameras, heavy 4K/8K models, and devices with edge analytics that push multiple video streams plus piles of metadata. For them, 1 Gbps is a genuine requirement, not bragging rights.

The difference between a consumer mindset and an engineering mindset is simple.

The consumer version: “If there’s gigabit, turn it on everywhere so there are no bottlenecks.”

The engineering version: “First, let’s see where the real bottlenecks are and only give gigabit there. Everything else stays on stable 100 Mbps, so we don’t create adventures for ourselves.”

To the camera, 100 Mbps is perfectly fine, especially if the routes are messy, the cable is old, and installation was done by someone who uses faith instead of a tester. But between switches, the NVR, the server, and storage — that’s where saving on gigabit quickly turns into archive problems and complaints like “everything lags when we try to view recordings.”

If you boil it down to a very down-to-earth rule, it’s this: don’t ask “can we make this gigabit?” Ask “what breaks if we leave this at 100 Mbps — and what do we actually gain by giving it 1 Gbps?”

In most cases, cameras live happily on honest 100 Mbps and quietly mind their business. Servers, uplinks, and the network core, on the other hand, truly appreciate your gigabits. In video surveillance, gigabit is not a “gold star for effort.” It’s a tool. In skilled hands, it makes the system fast and scalable. In the hands of someone who sets everything to “max,” it turns the network into an experimental lab where every camera is a lottery ticket.