How to plan a network for installing surveillance cameras at home? Wired or wireless, whether to install POE, I only figured out after installing six heads

Last year, my mom fell at home alone and lay down for over half an hour before she reached for the phone. After this incident, I made up my mind to install surveillance cameras. At first, my thoughts were simple: just buy a few cameras and connect them to WiFi, and that would be it. But in the first month, I had to redo it twice. The wireless model was offline every few days, and the battery model was too lazy to charge after two months and became just a decoration. After two years of struggling, now my home has six wired cameras, and I've made way more pitfalls than I imagined. This article goes through everything about network planning from start to finish. If you're planning to pretend, just copy it.

Let's do

the math: surveillance relies on uplink bandwidth, not download speed

many people do speed tests before installing surveillance, and when they see "300MB download," they think it's secure. Wrong, the camera transmits data outward; it feeds on uplink traffic. Home broadband's uplink is generally only a fraction of the downlink. My 500 Mbps broadband is rated at 30 Mbps upstream, but actual measurements are around 26 Mbps. A 4-megapixel camera running H.265 video recording delivers about 4 to 6 Mbps of bits per stream, with six heads uploading at full speed at 30 Mbps, which completely drains me up. The manifestation was: the surveillance footage looked normal, but my video call got stuck as a PowerPoint slide, because the video call data was squeezed into the upstream queue together.

so before installing, do one thing: go to speed testing websites and only check the upload section, then count how many heads you plan to install. For 30 Mbps uplink, don't exceed 4 million H.265 heads per 4 million heads; 2 million heads can hang a few more heads at 2 to 4 Mbps all the way. Some people say my family only has two or three heads, so whatever? Indeed, under three people, there's basically no need to calculate that amount. There's also a way to cut corners: manually limit the camera stream in the app to 2MB, so the image quality doesn't differ much to the naked eye, and the uplink pressure is cut in half. The signal from my backyard hanging outside the wall is average, so I just limited it to 1.5MB, and it's stable.

Wireless Cameras: Truly convenient, but also full of pitfalls

the first batch I bought was all wireless models, just because I didn't need to pull a cable. Only after using it did I realize the three pitfalls of this kind of thing. The first pitfall is network setup: most home cameras only support 2.4G WiFi and do not support 5G. If your router is connected to dual-band integration and your phone connects 5G to connect to the camera, eight out of ten times it fails because the camera can't hear the 5G broadcast. The first time I used a Xiaomi outdoor headset, it took forty minutes, and finally I had to temporarily turn off 5G in the router's backend before connecting. After installation, remember to disassemble the dual-band integration and leave a separate name for 2.4G; otherwise, if the router restarts, it might drift and disconnect again.

second pitfall is signaling. The camera is usually installed in corners like the door, backyard, or garage—these are the areas with the weakest WiFi in the home. On the backyard wall, I have a header on the backyard wall, a load-bearing wall and a yard gate on the wall. The signal is always one and a half bars, and the screen drops every few hours. The app taps in and spins for a long time before it shows up. The third pitfall is battery models: the advertiser claims "180 days of battery life," meaning sensor mode and shooting a dozen or so videos a day. The battery at my doorstep was running on for video all day, charging every three weeks. When I took it off and charged it, there was a vacuum surveillance at the door. Three months later, I downgraded it to a pure doorbell function. My conclusion for the wireless model is: it works next to indoor outlets or near the router, but don't rely on it in key locations like the doorway or backyard.

Wired plus POE: One network cable solves both power and the network

rework, I made up my mind to switch to wired and casually researched POE, and now I push it to everyone. POE simply means the switch powers the camera via a network cable. With just one cable, you have both network and power, so you don't need to leave a socket next to the camera or worry about charging. The specific purchase method is simple: choose the "POE model" for the camera (the product page will mention POE power, which is 20 to 30 yuan more expensive than the same wireless model), then buy a switch with a POE power port. I bought a TPLINK 8-port switch, with 4 ports with POE for 160 yuan, and 6 connectors for a bit more than enough (if you have enough ports, don't be greedy; POE has a total power cap. My one is labeled at 80 watts, 6 connectors per connector for a few watts, very generous).

cabling is the only challenging part of this solution. When I was renovating, I embedded network cables between the weak current box and each point, which made it so comfortable. If you haven't renovated before, you can use exposed wires, and white flat wires run along the baseboard, which is actually less noticeable. Network cables must be above Category 5, with each cable exceeding 80 meters. When installing, pay attention to two details: First, connect the POE camera to a separate switch to see if it can be found in the app, confirm it's good or bad, then climb the ladder. Once, after installation, I found the connector was broken, so I spent the whole afternoon disassembling and assembling it; Second, the switch is placed in a weak current box. In summer, feel the temperature and feel hot, leaving a gap in the box door. The network planning is the same idea as my previous article on renovation network cable planning: plan the locations in advance, lay them out in one go, and only then add equipment later to make choices.

Where is the best

for storing videos: SD card, cloud storage, NVR, NAS, I've paid for all four

cameras are installed, where to store the recordings is an even more costly issue than the network itself. I've tried all four options, so I'll give you a straight conclusion. SD cards are the cheapest way to start. Inserting a 64GB card into one head can store continuous recording for one to two weeks, costing fifty to sixty yuan. The downside is that if the camera is stolen, the recording is lost along with it. Also, it's normal for the card to fail once every year or so. If it breaks, you won't get any notification. I suffered this loss and only found out three months later that the card was already broken. Cloud storage is the most worry-free. You can review your phone anytime, but charge per head—over 100 yuan per year. My family has six kids in the first year, which is nearly 1,000 yuan. After three years, that's enough to buy a VCR. My compromise is: buy cloud storage on the front end (if lost, there's cloud backup), and use local storage for the rest.

the proper solution for local multiplexing is NVR recorders or NAS. An NVR is a dedicated recorder with a surveillance hard drive. Four or eight cameras start at two or three hundred yuan, plus a 2TB monitor just over 400 yuan. Storing whole-house recording for two or three months is no problem—the most worry-free family solution. Since I have a NAS at home, I store the recordings directly on the NAS: Feiniu installed a monitoring kit, set out an old drive just for recording, and automatically overwrite the old one when full. That drive was added according to the process I wrote before mounting the second hard drive the same setup. The advantage of a NAS solution is that you can review the phone app directly and do other tasks at once. The downside is you need some basic setup skills, making it easier for beginners and NVRs. The table below shows the costs I incurred. Please match according to your situation:

too many cameras, don't cram them all into the main router: set up a separate stall

Finally, let's talk about network structure—this is the one I most want to know in advance. Six cameras plus my phone, computer, and TV are all mounted under the main router. First, they take up the router's connection count, and second, the monitoring traffic will crowd out normal internet access. My current approach is to set aside a separate setup for the surveillance: the optical modem comes out and connects to the main router first, the main router's LAN port connects to the POE switch, and all cameras only connect to the switch. This means the surveillance setup only occupies "one device slot" under the main router, not six. I previously wrote an article using routers as switches the same idea: multi-port devices should be down-to-earth and don't let everything become a bottleneck. If you have many smart devices at home and old routers, this problem becomes even more obvious. I've done the math once for the load-carrying count, so if you're interested, check out this .

more detailed, you can assign a separate network segment to the surveillance (for example, the main router is 192.168.31.1, and devices under the monitoring switch go through another segment). The camera doesn't touch the internal network, so even if a firmware vulnerability is scanned, it won't reach your NAS or computer. This is a plus; beginners can ignore it for now. If you have few devices, a default network segment is sufficient. But here's one thing beginners should remember: the first thing you do after getting a camera is to change the default password. Many cameras have default passwords that can be found across the entire internet. If you don't change them, it's like leaving the door open for people to watch live streams. This is the most easily overlooked yet most important thing for those installing surveillance.

Copying Tasks According to Budget: Three Sets of Plans Directly Matched to Place

finally give three sets of combinations based on spending less, then copy them directly. The first set is the "100-yuan trial version": one or two wireless cameras plugged indoors, inserted into an SD card, total investment under 200 yuan, first test if you really need surveillance. I recommend everyone start with this set. The second set is the "reliable main version": POE cables at key locations like the doorway and backyard. Buy 8 POE switches plus NVR and 2TB monitoring drives, with four sockets plus a full set around 1,500 yuan, allowing for several months of storage. This is the optimal solution for most households. The third set, the "NAS All-in-One Edition": If you have a NAS at home, connect all the cameras to POE, record video into the NAS, and get a dedicated drive. Just follow the idea I wrote earlier. In total, you only spend a few more heads than the second setup.

my complete path was to upgrade from the first to the third, with the main tuition fees paid in between: battery fees and annual cloud storage fees. Looking back, if I had known from the start to install six heads, I could have jumped straight to the second set and saved nearly a thousand yuan. Let me summarize the order in one sentence: count the heads first, then count the upline; use wireless cables for socket edges; use POE cables at key locations; local recording is the main one; cloud only provides the entrance; all cameras are connected to a switch without pressing the main router; finally, change the default password. By following these steps, the surveillance system can run smoothly for many years.