WestPoint · Networks · Ubiquiti
It does a great deal for what it costs. And it has a ceiling.
Switches, Wi-Fi and outdoor links with a controller of its own that you understand in an afternoon. It is a very reasonable buy in a medium installation, and there are places where it falls short. Both sides are told here.
What kind of manufacturer it is, and what it is recognised for.
Ubiquiti — Wi-Fi, switching and outdoor links with a controller of its own. Its virtue is a ratio between price and capability that almost nobody else has.
Ubiquiti puts first that the network should be easy to see and easy to administer, at a price that does not look like the price of the big brands. Everything —switches, Wi-Fi points, router, outdoor links and its own cameras too— is governed from one controller with a clear screen, with maps, with consumption graphs and with the name of every connected device. Anyone who sees it for the first time understands their network better than before, and that has real value.
What it sacrifices is depth and backing. There is no full command line to get to the bottom of a problem with, the advanced network functions are the bare minimum, and the manufacturer's support does not work like that of a top-range house: it is resolved with documentation, community and equipment replacement. New versions sometimes change things that worked, and that is why here you do not update the day one comes out.
It is recognised in medium installations that are well built: an industrial unit with offices, a hotel, a large residential block, an industrial estate, a sports club. Places with dozens of points, not hundreds, and with somebody sharp inside but no network department.
How it is administered.
The controller of its own is half the product, and it is worth understanding where it lives.
There is an application that discovers all the devices, adopts them and configures them: the separated networks are created, it is decided which port goes to which network, you see how much power each port is giving and you see what device is hanging off it. For a camera installation this is very convenient: the map of which camera is on which port —which with other brands has to be built by hand— here comes out on its own.
That controller can be kept in the house, on a small device from the manufacturer itself or on a server, or you can use the manufacturer's cloud route to reach it from outside. The difference from a purely cloud brand is important and has to be said: if the controller is switched off, the network KEEPS working. The switches switch and the Wi-Fi gives service; what is lost is the administration and the statistics. The controller is not the brain of the traffic, it is the dashboard.
Where it fits well in a security installation.
In the access layer of medium installations, which is most of them: one cabinet or three, dozens of cameras, a recorder, traffic separation done properly and a real budget. There it does exactly what has to be done and leaves money to spend where it really matters, which is usually the camera, the drive and the cabling.
Where it falls short, and the brochure does not say it.
This is what you have to know before putting it in a large installation.
In the centre of a large network. When hundreds of cameras concentrate throughput towards a group of recorders, the work is fine-tuning: how much packet memory each unit has to take the traffic hits, how the queues are handled, how multicast is controlled port by port, how several networks are routed between each other with fine filters. There the functions exist in a basic version and the tool does not let you get down to the detail, and what cannot be measured cannot be fixed.
In fast recovery. Redundant paths are closed with the standard mechanisms, which take of the order of seconds to rebuild the network, and there are none of the almost instant ring protocols of industrial equipment, nor the behaviour of two units presenting themselves as one. If the installation cannot go blind even for a few seconds, this is not the range.
In support and in life cycle. A critical installation also buys somebody to call with a contract, a predictable version calendar and the guarantee of finding the same model in four years' time. Those three things are the weak point, and in an installation of fifty cameras they are hardly noticed; in one of five hundred, they are all noticed at once.
What really decides a video project.
Four things. The first has crossed our path in other people's installations more often than any other.
- The power, which is where the low price collects. There are models with many PoE ports and a modest total budget, and the mixture is dangerous: it looks as if twenty-four cameras fit because there are twenty-four ports. You add up each camera's winter consumption, compare it with the unit's total and leave a third as margin. You also have to look at which power standard each port supports, because cameras with a heater or with strong illumination ask for more than a basic port gives.
- Multicast, if the video system uses it. When several screens watch the same camera, some systems send a single copy that the network shares out. If the switch is not configured to listen for who wants that copy, it sends it to every port and one single stream saturates the whole network. Here the necessary control can be switched on, but it has to be switched on and you have to check that somebody on that network is asking who wants what: the classic failure is that it works for two minutes and then stops.
- The uplink path and the packet memory. A cheap access switch with a single link to the main cabinet and little memory drops packets when the bursts from several cameras coincide. It does not show on any average-usage graph: it shows in the recording, as broken pieces of image. You spread the load across more links or you move up a range at the concentration point, which is usually a single unit.
Licences and subscriptions, and what happens when they expire.
The network part is not on subscription: the equipment is bought, the controller is not paid for and neither are the updates. This is part of why the sum works out, and it is worth saying plainly because on this same list there are brands where the licence is what keeps the equipment switched on. Here, no: what is bought works even if nobody pays anything more.
Who administers it when we leave.
It is one of the brands where the customer really does keep their network, and that is an argument in its favour.
The day to day is run by the customer without argument: seeing which camera is disconnected, restarting a port, adding a Wi-Fi point, changing the guest network password, looking at how much traffic there is. The screen is meant for that and a two-hour session with a one-page procedure is enough. In medium installations it is the brand with which the most customers end up administering their network themselves.
What stays on our side, or on the side of whoever maintains it: the design of the traffic separation and of the addressing, the rules on what can talk to what, the power calculation when cameras are added and the updates. That last one is the important one: it is the brand where an automatic upgrade has changed a behaviour most often, so it is better that somebody decides it and not the calendar.
And if the customer's IT team already has its brand for the whole house, we are not coming to change it. Very often the right conversation is the opposite one: they command their network, we put in the camera part with whatever they ask of us, and if what they ask of us does not reach what is needed, it is explained with the sum done and before signing.
What we do, and what we do not promise.
The power and throughput sum before choosing a model, the separation of the cameras with the way out to the internet closed, the controller inside the house and backed up, the cameras with fixed and documented addresses, and the firmware version fixed and noted down. With the usual tests: everything switched on while measuring watts, the uplink saturated on purpose and a cable cut with a stopwatch.
We are not an authorised channel or a certified integrator for anyone, and we are not going to tell you this is a better brand than another. We will tell you how far it goes and what happens beyond that.
Where to go next.
What does not change with the brand —how much throughput the video generates, why the cameras go separated, when fibre is called for— is told in the area.
Get started
A medium installation and a real budget?
Tell us how many cameras there are, of what kind, in how many cabinets and whether there is somebody inside who wants to run the network. We will tell you what is solved well with this range, which point is worth reinforcing and where to spend the money you save.