How to Secure IoT Devices: A Practical Guide
Securing an IoT device comes down to five controls, done in this order: change its default password, put it on a separate network from your computers, keep its firmware updated, turn off the features and remote access you do not use, and watch what it talks to. No single product does this for you, and no device is safe because it is small or cheap — the largest botnet attack in internet history was built from cameras and routers exactly like the ones in homes and small offices.
IoT means the Internet of Things: the cameras, smart locks, thermostats, doorbells, printers, smart TVs, voice assistants, baby monitors, and sensors that connect to your network. A laptop gets antivirus software, regular patches, and a user who notices when something is wrong. An IoT device usually gets none of that. Most of them run stripped-down software you cannot inspect, cannot install security tools on, and often cannot update at all. That is the whole problem in one sentence: you have to protect these devices from the network side, because you cannot protect them from the device side.
Why IoT security is harder than laptop security
Four differences explain almost every IoT breach.
You cannot install protection on the device. There is no antivirus agent for a smart plug. The manufacturer controls the software, and on many cheap devices the software is never updated after the product ships. If a flaw is found two years later, the device stays vulnerable forever.
Default credentials are the norm, not the exception. Cameras, routers, and recorders have shipped for years with usernames like admin and passwords like admin or 12345, identical on every unit. Attackers do not guess these — they look them up in public lists. The Mirai botnet of 2016 infected hundreds of thousands of cameras and routers this way, using a short list of factory default logins, and used them to take major websites offline with a distributed denial-of-service attack. The devices were not hacked in any clever sense. They were logged into.
Devices outlive their support. A laptop gets replaced every few years. A smart thermostat, a security camera, or a network printer can sit on a wall for a decade — long after the manufacturer has stopped publishing firmware updates. An unsupported device is a permanent open door: every flaw discovered after support ends stays open.
One compromised device reaches the rest of the network. A smart TV on the same network as your work laptop can be a stepping stone. Attackers use a weak device to get inside, then move sideways to the computers that hold real data. The camera was never the target. It was the door.
None of this means you should throw the devices out. It means the security work happens in a different place: at the password, at the network, and at the buying decision.
1. Change every default password first
This is the single highest-value action, and it costs nothing. Before a device goes on your network — or the moment you realise one already on it still has its factory login — change the username and password to something unique and strong, and store it in a password manager. Do this for the device itself and for the app account and cloud dashboard that manage it. If a device will not let you change its password, that is a serious finding: isolate it on its own network segment (step 2) or replace it.
Use a different password per device. Reusing one password across ten devices means one leaked credential opens all ten. Enable multi-factor authentication on the management account wherever it is offered — email, cloud dashboard, and the manufacturer’s app. A camera whose cloud account has no second factor is one phished password away from being watched by a stranger.
2. Put IoT devices on their own network
This is the single highest-impact technical control, and it is the one most homes and small offices skip. Put all IoT devices on a separate Wi-Fi network or VLAN from your computers and phones, with firewall rules that let each device reach only what it genuinely needs — usually its manufacturer’s cloud service — and nothing else. A camera should never be able to reach your file server, your finance computer, or another camera.
Most modern routers make this easy: use the guest network, or create a dedicated IoT SSID, and put the devices there. The practical test is simple. After segmenting, ask: if this camera were fully compromised tomorrow, what could the attacker reach? The answer should be “other cameras and the internet” — never your laptop, your backups, or your work systems.
Segmentation also limits the blast radius of the one device you forgot about. The old printer in the corner, the smart plug nobody remembers installing — on a segmented network, they are contained. On your main network, they are a bridge.
3. Keep firmware updated — and track support end dates
Check for firmware updates when you install a device, then check again on a schedule — monthly is enough for a home, and a named person should own it in a small business. Turn on automatic updates wherever the device offers them. An update is the only way a flaw in the device’s software ever gets fixed, and unpatched devices are the second most common way in after default passwords.
Just as important is the support question most people never ask: how long will this manufacturer keep updating this device? Before you buy, look for a published support end date. If the manufacturer does not publish one, treat that as an answer. A device with two years of updates left is a different purchase from one with five. When support ends, the honest options are to replace the device or to isolate it so tightly that its unpatched flaws cannot be reached — not to leave it on the main network and hope.
4. Turn off what you do not use
Every open port, enabled service, and remote-access feature is a door. Disable remote management unless you genuinely need it, and if you do need it, reach the device through a VPN rather than exposing its management page to the internet. Turn off UPnP on your router — it lets devices open ports to the internet by themselves, which is exactly the convenience attackers exploit. Disable voice control, cloud recording, or app features you never use; each one is a connection to a server you do not control.
Check the device’s settings for a simple question: does this need to be reachable from outside my home or office? For most devices the answer is no. A thermostat does not need an open port. A camera that you only view from home does not need cloud access at all.
5. Watch what the devices talk to
You cannot install monitoring software on a camera, but you can watch its traffic from the router. Learn what normal looks like — a camera talks to its manufacturer’s servers, a thermostat checks in occasionally — and notice deviations: a camera suddenly sending large amounts of data to an unfamiliar address, a sensor generating far more traffic than usual, a device trying connections it has never made before. Many routers show per-device traffic; a few minutes a month scanning that list catches the device that has gone wrong.
This is also how you find the devices you forgot. A quarterly network scan — your router’s connected-devices list, or a simple network scanner — turns up the old plug, the retired camera, the printer nobody uses. Every unknown device is a device with an unknown password and unknown firmware. Either claim it, update it, and segment it, or take it off the network.
Before you buy: five questions that prevent most problems
Most IoT security failures are decided at purchase, not at setup. Before buying any connected device, ask:
- Can I change its password, and does every unit ship with a unique one? If the answer is no, do not buy it.
- How long will the manufacturer publish security updates, and is there a published end date? No answer means short support.
- Does it support encrypted connections, and does its app offer multi-factor authentication? Both should be yes in 2026.
- Can it work without cloud access, or does it stop working if the manufacturer’s servers go away? Local control is a security feature, not just a convenience.
- Is there a way to report a vulnerability to the manufacturer? A company with no security contact is a company that does not fix flaws.
A cheap device that fails three of these is not a bargain. It is a device you will be securing, or replacing, for years.
When a device has to go
Before you sell, give away, or throw out a smart device, factory-reset it. Cameras hold Wi-Fi passwords. Voice assistants hold account tokens. Smart locks hold access codes. A reset is the only reliable way to remove them, and it takes two minutes. The same applies to devices you stop using: a retired camera left powered on and connected is still a network device with a password you have forgotten. Unplug it, reset it, or remove it from the network entirely.
The priority order, if you do nothing else
If you have an afternoon and a house full of devices, work in this order: change the default passwords, move the devices onto a separate network, check for firmware updates, and turn off remote access you do not use. Those four steps, done once and maintained lightly, put you ahead of the overwhelming majority of homes and small offices — because most IoT breaches still start with a default password on a device sitting on the same network as everything that matters.
FAQ
What is the biggest IoT security risk?
Default passwords. Most IoT attacks still start with factory credentials that were never changed — the Mirai botnet infected hundreds of thousands of cameras and routers this way. Changing every default password to a unique, strong one is the highest-value step you can take.
Should IoT devices be on a separate Wi-Fi network?
Yes. Put them on a guest network or a dedicated IoT network, separate from your computers and phones, with firewall rules that let each device reach only what it needs. That way a compromised camera cannot reach your laptop, your files, or your backups.
How do I know if an IoT device is still supported?
Check the manufacturer’s website for a published support or security-update end date. If there is none, contact support and ask. When support ends, the device receives no more fixes — replace it, or isolate it so tightly that its flaws cannot be reached from the rest of your network.
Can a smart TV or camera really be hacked?
Yes — and the device itself is rarely the target. Attackers use a weak device as a way into the network, then move to the computers that hold real data. That is why network segmentation matters more than the security of any single device.
Do I need antivirus for IoT devices?
You cannot install antivirus on most IoT devices — they run closed software with no room for security tools. Protection has to happen at the network instead: strong unique passwords, segmentation, firmware updates, and watching the device’s traffic from your router.
