The address on the sticker is a fallback, not a promise. On a home router it is usually right — the thing is a DHCP server sitting on a fixed LAN address, and a factory reset puts it back there. On almost everything else built in the last fifteen years, the documented address only appears when nothing on the wire answers a DHCP request. Plug it into a working network and it takes a lease instead; the number in the manual is never used.
- Most common gateway addresses
- 192.168.1.1 and 192.168.0.1
- Cable modem diagnostics
- 192.168.100.1, present in bridge mode too
- No DHCP server on the wire
- Vendor static fallback, or link-local 169.254.0.0/16
- Typical fallback delay
- 60 to 120 seconds after link-up
- Web management ports
- 80 and 443; 8080 and 8443 as alternates
- Definitive answer for a router
- Your own default gateway
Why the documented address is often not the address
There are two populations of network device and they behave in opposite ways.
The first is the gateway: a home or small-office router, an ISP hub, a firewall. It owns the subnet and runs the DHCP server, so it has to sit at a known address before anything else on the network exists, and that address is baked into the factory configuration. Here the documented address is reliable — provided the unit has been reset, you are in a LAN port rather than the WAN port, and nobody has re-addressed it since.
The second is everything else: access points, managed switches, IP cameras, NVRs, baseboard management controllers, door controllers, desk phones, print servers. These ship as DHCP clients. The address printed in the quick-start guide is what the device falls back to when no DHCP server answers, which on a working network is never. That is the single largest reason people type 192.168.1.64 into a browser and get nothing.
Four other things routinely invalidate a documented default.
- The vendor changed it between generations. ASUS moved from 192.168.1.1 to 192.168.50.1 on the ZenWiFi and recent Wi-Fi 6 lines. TP-Link uses 192.168.0.1 on some lines, 192.168.1.1 on others and 192.168.68.1 on every Deco. Axis dropped its static default altogether at AXIS OS 12.0: devices on 11.x and earlier sit on 192.168.0.90 when DHCP fails, and newer ones have no default address at all — a lease, or a link-local 169.254.x.x, and nothing else.
- The ISP flashed its own firmware. A BT Smart Hub answers on 192.168.1.254, not 192.168.1.1. Sky and Virgin Media hubs use 192.168.0.1. Put a Virgin hub into modem mode and its management address moves to 192.168.100.1, because at that point it is behaving as a cable modem rather than a router.
- Somebody changed it. A reboot does not undo a configuration change; only a reset to factory defaults does, and on most devices that means holding the recessed button for a specified count of seconds while the unit is powered, not while it is booting.
- The address exists but is not routed from where you are standing. This one catches experienced people.
That last case deserves a picture. A cable modem keeps its 192.168.100.1 diagnostic server alive even while bridged, and that address is near-universal across DOCSIS hardware — but it lives on the modem’s own management segment, so a laptop at 192.168.1.50 has no route to it. A recorder with built-in PoE ports does the same thing: a Hikvision NVR runs a second interface at 192.168.254.1 and re-addresses attached cameras from 192.168.254.2 upwards, Dahua uses 10.1.1.1. The cameras have addresses. You cannot reach them.
Why tplinkwifi.net fails
Hostname shortcuts — tplinkwifi.net, routerlogin.net, router.asus.com, fritz.box, myrouter.local — are answered by the device’s own DNS responder, not by the public DNS system. Switch on DNS over HTTPS in the browser, or point the machine at 1.1.1.1 directly, and the query never reaches the router: the name fails while the IP address works perfectly. The hostname is a convenience, never a diagnostic.
DHCP client first, static fallback second
On link-up the device broadcasts DHCPDISCOVER and waits. If an offer arrives it takes the lease, and the documented address is never used. If nothing answers inside the vendor’s timeout — usually a minute or two — one of three things happens. Some devices adopt the documented static address. Some take a link-local address from 169.254.0.0/16 under RFC 3927, picking one pseudo-randomly and ARP-probing to check nobody else has it. And a third group does both: it adopts the fallback and keeps sending DHCPDISCOVER in the background, so the address you find it on is temporary.
Between vendors the behaviour varies more than the addresses do.
- Cisco Business switches take 192.168.1.254/24 when no lease arrives and carry on discovering. The fallback address is dropped the moment a lease or a manual address is configured on any VLAN.
- Netgear smart managed switches fall back to 192.168.0.239 once the DHCP attempt times out, which takes about a minute.
- D-Link managed switches run no DHCP client at all out of the box. They sit on 10.90.90.90 with a 255.0.0.0 mask and stay there until you tell them otherwise.
- Lenovo XClarity Controller names the mode in its own interface — “first DHCP, then static IP” — with 192.168.70.125 as the static value.
- UniFi access points and switches fall back to 192.168.1.20, which the controller treats as reserved. Keep your DHCP pool clear of it, or you will eventually hand that address to a printer and spend an afternoon on it.
Connect a device straight to a laptop and two things follow from that timer. You have to wait it out — plenty of people try for twenty seconds and conclude the unit is dead. And your laptop is now a DHCP client with nothing to talk to, so it hands itself a 169.254 address on a different subnet from the device.
- Isolate the device. One cable, laptop to device. No switch in between that might carry a DHCP server, and no Wi-Fi on the laptop — a second default route makes the routing decision ambiguous.
- Give the laptop a static address in the fallback subnet. For a Netgear switch at 192.168.0.239, use 192.168.0.210/24. For a Cisco Business switch at 192.168.1.254, use 192.168.1.10/24. For a D-Link switch at 10.90.90.90, use 10.90.90.100/8. Gateway can be blank.
- Power-cycle the device and wait. Two minutes from link-up, not twenty seconds.
- Confirm at layer 2 before layer 3. Ping the fallback address once, then run
arp -a. An entry with a real MAC proves the device is there and answering even if HTTP is not. No ARP entry means the address is wrong, not the port. - Try HTTPS as well as HTTP. Recent firmware on switches and BMCs frequently drops plain HTTP, and a refused connection on port 80 looks identical to a wrong address.
Before you reach for the reset button
A factory reset on a switch, firewall or recorder destroys VLANs, port configuration, camera bindings and recorded-stream settings, and on some platforms it also clears licence activation state. It is the last step in this article, not the first. Exhaust discovery, serial console and the DHCP lease table before you hold the pin in.
Addresses by manufacturer
Two tables, because the addresses mean different things. In the first, that is where the device lives. In the second, it is where it lives only when nothing answers.
| Manufacturer | Line | Address | Hostname shortcut |
|---|---|---|---|
| Netgear | Nighthawk, most routers | 192.168.1.1 | routerlogin.net |
| Netgear | Orbi | 192.168.1.1 | orbilogin.com |
| TP-Link | Archer and most routers | 192.168.0.1 or 192.168.1.1 | tplinkwifi.net |
| TP-Link | Deco mesh | 192.168.68.1 | tplinkdeco.net |
| ASUS | Older RT models | 192.168.1.1 | router.asus.com |
| ASUS | ZenWiFi, recent Wi-Fi 6/6E | 192.168.50.1 | router.asus.com |
| Linksys | Most routers | 192.168.1.1 | myrouter.local |
| D-Link | Routers | 192.168.0.1 | dlinkrouter.local |
| Belkin | Routers | 192.168.2.1 | http://router |
| AVM | FRITZ!Box | 192.168.178.1 | fritz.box |
| MikroTik | RouterBOARD default config | 192.168.88.1 | — |
| Xiaomi | Mi Router | 192.168.31.1 | miwifi.com |
| Huawei | Mobile and LTE units | 192.168.8.1 | — |
| Huawei | HG-series ONT | 192.168.100.1 | — |
| Zyxel, DrayTek, ZTE | Most routers | 192.168.1.1 | — |
| Tenda | Routers | 192.168.0.1 | tendawifi.com |
| Nest Wifi, Google Wifi | 192.168.86.1 | Google Home app; no web UI | |
| Apple | AirPort base stations | 10.0.1.1 | AirPort Utility |
| Comcast / Xfinity | Supplied gateways | 10.0.0.1 | — |
| BT | Home Hub, Smart Hub | 192.168.1.254 | bthomehub.home |
| Sky | Broadband Hub | 192.168.0.1 | — |
| Virgin Media | Hub, router mode | 192.168.0.1 | — |
| Virgin Media | Hub, modem mode | 192.168.100.1 | — |
| Arris, Motorola, Technicolor | DOCSIS cable modems | 192.168.100.1 | diagnostics, read-only |
| Ubiquiti | UniFi gateways, Dream Machine | 192.168.1.1 | unifi.ui.com, via the cloud |
| OpenWrt, pfSense, OPNsense | Default install | 192.168.1.1 | — |
| Manufacturer | Device class | Fallback address | Behaviour |
|---|---|---|---|
| Hikvision | Cameras, NVRs | 192.168.1.64 (192.0.0.64 on some models) | DHCP on by default; find with SADP |
| Hikvision | NVR internal PoE ports | 192.168.254.1, cameras from .2 up | Separate interface, not routed from the LAN |
| Dahua | Cameras, NVRs | 192.168.1.108 | DHCP on by default; find with ConfigTool |
| Dahua | NVR internal PoE ports | 10.1.1.1, cameras on 10.1.1.0/24 | Separate interface |
| Uniview | Cameras, NVRs | 192.168.0.13 or 192.168.1.13 | Varies by model; find with EZTools |
| Axis | AXIS OS 11.x and earlier | 192.168.0.90 | Static fallback |
| Axis | AXIS OS 12.0 and later | None | DHCP, else link-local; use AXIS IP Utility |
| Ubiquiti | UniFi APs and switches | 192.168.1.20 | Reserved by the controller; keep DHCP pools clear of it |
| Cisco | Business and Small Business switches | 192.168.1.254/24 | Adopted after the DHCP timeout, keeps discovering |
| Cisco | Catalyst, ISR, IOS platforms | None | Console only until an SVI is configured |
| Netgear | Smart and managed switches | 192.168.0.239 | Adopted about a minute after link-up |
| D-Link | Managed switches | 10.90.90.90 / 255.0.0.0 | Static, no DHCP client by default |
| Dell | iDRAC9 | 192.168.0.120 | DHCP is the shipped default; reset returns to DHCP |
| Lenovo | XClarity Controller, XCC2 | 192.168.70.125 | “First DHCP, then static IP” |
| HPE | iLO 5, iLO 6 | None | DHCP only; set it in the iLO configuration utility at POST |
| Aruba | Instant APs | 172.31.98.1 | Reached over the SetMeUp-xx:xx:xx provisioning SSID |
Where a brand is missing, the per-manufacturer pages carry model-level detail alongside the shipped credentials — start at the A–Z manufacturer index, or go straight to a common one such as Netgear, Hikvision or MikroTik.
MikroTik is a special case
RouterBOARD units ship with a default configuration that puts 192.168.88.1/24 on the LAN bridge, but a device reset with the “no default configuration” option has no IP address anywhere. It is still reachable: Winbox and mac-telnet connect to the MAC address over layer 2, and RouterOS announces itself with MikroTik Neighbor Discovery on UDP 5678. No IP required at either end.
Management ports and what they imply
An address on its own is half the answer. What is listening tells you which management path the device expects you to use, and on a device you did not deploy, something about how it was configured.
| Port | Service | What it implies |
|---|---|---|
| 80 / 443 | HTTP / HTTPS | The primary web UI. Newer firmware often disables 80 entirely or redirects it |
| 8080 / 8443 | Alternate web UI | Something already owns 80/443 — a controller application, or the device is behind a port-forward |
| 22 | SSH | A real CLI exists. On switches and firewalls this is usually the faster route |
| 23 | Telnet | Legacy. Off by default on current firmware; open on a WAN interface it is a finding, not a feature |
| 161/udp | SNMP | Answers to a community string. Read-only public is still depressingly common |
| 623/udp | IPMI RMCP+ | A baseboard management controller is present even if its web UI is filtered |
| 7547 | TR-069 / CWMP | ISP remote management. Listens on the WAN side by design, so the ISP’s ACS can push configuration |
| 8291 | Winbox | The device is running RouterOS |
| 5678/udp | MikroTik MNDP | Neighbour discovery; works without an IP address |
| 10001/udp | Ubiquiti discovery | UniFi or airOS hardware on the segment |
| 5353/udp | mDNS | The device advertises a .local name you can use instead of an address |
| 8000 | Hikvision service port | The mobile app and SDK path, separate from the web UI on 80 |
| 37777 / 37778 | Dahua TCP / UDP | Native client protocol |
| 554 | RTSP | Video streams, not management. Its presence confirms a camera or recorder |
Port 7547 explains a whole class of “my router changed on its own” reports. TR-069 puts a small HTTP listener on the customer device, and the ISP’s auto-configuration server connects to it to push firmware, rewrite settings and read state with nobody touching the hardware. If your ISP hub changed its LAN subnet or re-enabled something you had turned off, that is usually how. Sitting on the ISP-facing side of hundreds of millions of units, it has also been mass-exploited more than once.
On a device you control, a sane configuration is narrow: management on 443, SSH on 22 if you need a CLI, SNMP either disabled or bound to a management VLAN with a v3 credential, and nothing at all reachable from the WAN interface.
Finding the address when none of these work
Work down this list. It is ordered by how much it costs you, not by how clever it is.
- Read your own default gateway. For a router this is definitive and takes five seconds. Windows:
ipconfig. Linux:ip route show default. macOS:ipconfig getoption en0 router. Whatever it prints is the router, regardless of what the manual claims. - Check the DHCP server’s lease table. For anything that is a DHCP client this is the answer, and it gives you the MAC and often the hostname as well. On home routers it is listed under “attached devices” or “DHCP clients”.
- Sweep the segment and match the MAC. Run as root,
nmap -sn 192.168.1.0/24does an ARP sweep of the local subnet and finds devices that ignore ICMP. The MAC on the device label tells you which of forty unknown addresses is yours; the first three octets are the OUI, so they also identify the vendor. - Use the vendor’s discovery tool. These speak broadcast or multicast and find devices on the wrong subnet entirely, which is exactly the case you are in. Hikvision SADP, Dahua ConfigTool, Uniview EZTools, AXIS IP Utility, Ubiquiti’s device discovery, Winbox’s Neighbors tab.
- Browse mDNS.
dns-sd -B _http._tcpon macOS,avahi-browse -arton Linux. This is how you find a link-local device that never took a lease. - Ask the switch. On a managed switch,
show mac address-tabletells you which physical port the MAC is on, andshow lldp neighborsorshow cdp neighborswill often print the device’s management address directly. - Listen on a direct cable. The best trick in the list. Connect the device to your laptop and run
sudo tcpdump -n -i en0 'arp or port 67 or port 68'. It announces itself: the DHCPDISCOVER carries its MAC and often its hostname, and the gratuitous ARP it sends after falling back carries its address and subnet. You are reading the answer off the wire rather than guessing. - Serial console. The only route into a Cisco IOS device out of the box and the reliable route into most others. 9600 8N1 is the common default; MikroTik and Ubiquiti use 115200. Check the pinout before you plug an RJ45 console cable into an Ethernet port.
- Factory reset. Now the documented address in the tables above applies — and so does everything in the first section about why it might still not.
Finding a device that has no IP address covers the layer-2 approaches in more detail. And if you have the right address but the page will not open, that is a different fault: the router admin page not loading is usually HSTS, a cached redirect to HTTPS, or a proxy, not a wrong address.
When you do get in, put the address on the unit. A label with the management IP, the VLAN and the date beats any documentation that lives somewhere else.



