Scroll to turn the page
Networking · AP Networking
a home or small-office network
One internet connection, shared by every device in a home or a small office: that is a SOHO (small office/home office) network.
In the figure (hypothetical) a wireless gateway brings the internet in from the ISP and shares it. A switch joins the wired devices; 3 devices join without wires. While the phone loads a website, the desktop prints and the laptop saves a file to network storage, all at once, on one network.
With many devices sharing it, more can go wrong, and more must be planned: whose address is whose, which cable or connection to use, what to secure.
This is Unit 2 of AP Networking, second of four. Unit 1 managed one device’s connections; this unit manages a shared network: troubleshooting it (2.1), documenting it with its addresses (2.2), upgrading it (2.3), adding features such as a media server and checking an AI’s advice (2.4), and securing it (2.5, 2.6). Unit 3 goes on to networks with many more connections.
troubleshooting my SOHO network
Slow or no network connectivity is a common issue in a SOHO network. Any of its devices and physical connections can be the source.
Many issues are resolved by common solutions: restarting the network router or wireless access point; reseating (removing and reinserting) a loose or unplugged cable; replacing damaged cables or connectors. If that does not restore it, keep troubleshooting by gathering diagnostic information.
A hardware failure or a disconnection shows in the indicator lights on the router, switch or access point. Check the device’s manual or label to read them; typically:
| light | may indicate |
|---|---|
| steady or flashing green | normal operation |
| red or orange | a hardware issue or connection failure |
| unlit | a faulty port or cable, or an unplugged cable |
For red, orange or unlit lights, try a different Ethernet cable, or plug the cable into a different port. In the figure (blue stands for a green light) port 3 turns red and is fixed with a different Ethernet cable; port 4 goes dark and is fixed by reseating the cable.
response times and lost packets
Network congestion or a hardware failure can make applications, websites and services slow or unreachable. The command-line tool ping checks connectivity and response times.
ping 8.8.8.8 tests external connectivity: it sends requests to one of Google’s public servers and times each reply.
Response times over 100 ms or lost packets may indicate congestion or hardware issues. No response, or 100% packet loss, may indicate a local disconnection or failure.
In the figure (hypothetical times) a laptop on Wi-Fi, two rooms from the access point gets 3 of 4 replies, 25% loss, averaging 179 ms: congestion or a weak signal.
For response times over 100 ms or more than 1–2% packet loss, try a wired connection instead of wireless, pausing large downloads, or moving the device closer to the access point. Plugged in by Ethernet, the same laptop gets 4 of 4 replies averaging 12 ms.
APIPA and loopback addresses
An incorrect IP configuration can keep a device from communicating. Check it with ipconfig (Windows) or ifconfig (Linux).
An address in the range 169.254.0.0–169.254.255.255 is an APIPA (Automatic Private IP Addressing) address: there was a problem assigning the device an address. Fix it by disconnecting and reconnecting to the network, or by releasing and renewing the address: ipconfig /release, then ipconfig /renew.
In the figure (hypothetical) the laptop shows 169.254.87.12. After the release and renew, the DHCP server on the router gives it 192.168.1.23, mask 255.255.255.0, gateway 192.168.1.1.
Only the loopback address 127.0.0.1 means the device may not be connected to any network: it can send data only to itself. Make sure its Wi-Fi and network adapter are enabled, and restart it.
Then verify: the device reaches the internet or network resources (shared drives, printers, servers), and websites and services load without delay. If not, gather new diagnostic information and try further solutions.
documenting my network
A node is any device connected to a network. A host is a node that sends or receives data, such as a computer or a printer; hosts are often called endpoints. Switches, routers and access points are nodes but not hosts.
A device needs a network interface card (NIC) to connect. Each NIC has a MAC (media access control) address, also called a physical or hardware address. It is hardcoded into the NIC and stays the same when the device moves to another network.
A MAC address is a 48-bit identifier, written as 12 hexadecimal digits. The first half, the OUI (Organizationally Unique Identifier), names the manufacturer (blue in the figure); the second half names the device (red). The same address in its three formats:
| format | example |
|---|---|
| no separators | 6F72696F6E2A |
| colons | 6F:72:69:6F:6E:2A |
| hyphens | 6F-72-69-6F-6E-2A |
Hexadecimal is base 16, with the digits 0–9 and A–F. One hex digit stands for 4 bits, so it is far more compact than binary. To convert from decimal, divide by 16, keep the remainder as a hex digit, divide the quotient again, until the quotient is 0; read the digits last to first. 79 is 4F:
| step | divide | quotient | remainder | hex digit |
|---|---|---|---|---|
| 1 | 79 ÷ 16 | 4 | 15 | F |
| 2 | 4 ÷ 16 | 0 | 4 | 4 |
how a switch finds a device
A switch stores the MAC addresses of the devices physically connected to it in its content addressable memory (CAM) table.
The table says which port each MAC address is on, so the switch sends data only to the port of the device it is for, instead of broadcasting it to all ports. The framework’s example:
| device | MAC address | port |
|---|---|---|
| A | 6D:6F:72:69:61:68 | 1 |
| B | 4F:72:69:6F:6E:2A | 2 |
| C | 70:6F:6C:61:72:24 | 3 |
When device A sends data to device B, the switch looks up B’s MAC address and forwards the data only to port 2. Device C never sees it. In the figure A sends to B, then C sends to A, and each time one row of the table lights.
32 bits in four octets
An IP (Internet Protocol) address is a logical address, assigned to a device on a specific network. When the device moves to another network, its IP address typically changes; its MAC address does not.
An IPv4 address is a 32-bit number divided into four 8-bit sections called octets. It is written in dotted decimal: each octet as a decimal number from 0 to 255, separated by periods.
To convert an octet, add the place values (128, 64, 32, 16, 8, 4, 2, 1) of its 1 bits. The framework’s example, 10001110.11111010.10111111.01001110, is 142.250.191.78:
| octet | bits | decimal |
|---|---|---|
| 1 | 10001110 | 142 |
| 2 | 11111010 | 250 |
| 3 | 10111111 | 191 |
| 4 | 01001110 | 78 |
An IPv4 address has two parts: a network ID, shared by all devices on a local network, and a host ID, which identifies each device on it.
network bits and host bits
The network ID of an IPv4 address may be 1–3 octets, depending on the class of the network. The size of the host ID sets how many hosts the network can hold.
Classful addressing divides the IPv4 address space into five classes. Classes D and E are reserved for other purposes (multicast and experimental use).
| class | network ID | default mask | host addresses |
|---|---|---|---|
| A | 1 octet | 255.0.0.0 | 16,777,214 (about 17 million) |
| B | 2 octets | 255.255.0.0 | 65,534 (about 65,000) |
| C | 3 octets | 255.255.255.0 | 254 (about 254) |
Every IPv4 address is paired with a subnet mask, which marks the network bits (the mask’s 1s) and the host bits (its 0s). With 255.255.0.0 the first two octets are network bits and the last two are host bits; with 255.255.255.0 the fourth octet alone is for host addresses.
A host ID of h bits gives 2h − 2 host addresses: all 0s names the network itself and all 1s is its broadcast address. Class C: 28 − 2 = 254. Class B: 216 − 2 = 65,534. Class A: 224 − 2 = 16,777,214. By its first octet, a class A address starts 1–126, class B 128–191, class C 192–223.
types of IPv4 address
A public address is unique and assigned by the ISP. It identifies a device that is directly reachable from the internet and is used to communicate on the internet and with other networks.
Private addresses are used within a LAN (local area network) only: routers on the internet drop any data addressed to a private IP address. That adds a layer of security, since privately addressed devices cannot be accessed directly from outside.
| kind | range | used for |
|---|---|---|
| private, class A | 10.0.0.0–10.255.255.255 | inside a LAN only |
| private, class B | 172.16.0.0–172.31.255.255 | inside a LAN only |
| private, class C | 192.168.0.0–192.168.255.255 | inside a LAN only |
| APIPA | 169.254.0.0–169.254.255.255 | no DHCP server answered |
| loopback | 127.0.0.1 | the device itself, for testing |
| public | everything else in use | reachable from the internet |
APIPA addresses hosts automatically when the DHCP (Dynamic Host Configuration Protocol) server, which automates IP assignment and configuration, is unavailable or misconfigured. Loopback addresses let a device send data to itself for testing; 127.0.0.1 is the local machine.
In the figure (hypothetical addresses) the laptop’s request leaves through the router’s public address 203.0.113.7; data from the internet addressed to 192.168.1.23, a private address, is dropped; and the laptop pings 127.0.0.1, which never leaves it.
from an IP address to a MAC address
The Address Resolution Protocol (ARP) maps a device’s IP address to its MAC address, so that data are delivered to the right device.
To send to another device on the same local network, a device must know its MAC address. It sends an ARP request, holding the destination IP address it knows, to all devices on the network.
Only the device with that IP address answers, with an ARP reply that contains its MAC address.
In the figure (hypothetical addresses) the laptop, 192.168.1.23, asks everyone “who has 192.168.1.40?”. The printer answers “192.168.1.40 is at 00:80:77:5A:21:C3”, to the laptop alone. The desktop and the NAS ignore the request.
devices, addresses and connections
A network diagram plans a network before it is built and documents it after setup. It shows the routers, switches, access points, servers and endpoints, with solid lines for wired connections and dashed lines for wireless ones, and each device’s MAC and IP address, to help with troubleshooting and expansion.
A SOHO network serves a few users on a single LAN. It commonly includes routers, switches, access points, endpoints, firewalls and cabling:
| device | role |
|---|---|
| router | directs data between the local network and other networks; often combined with a modem, for internet access through an ISP |
| switch | connects wired devices and manages traffic within the LAN |
| wireless access point (WAP) | gives wireless connectivity, by connecting to a router or switch |
| wireless gateway | one device that combines a modem, router, switch and WAP |
| endpoints | laptops, PCs, printers, NAS, servers, 3D printers, VoIP phones, IoT devices, mobile phones |
Twisted pair (Ethernet) cable is the most common for internal wired connections. Coaxial and fiber optic cables are often used by ISPs to connect to the SOHO router from outside.
The figure’s diagram: 4 wired links, 3 wireless, and a coaxial line from the ISP. Click the gateway to break it into its 4 parts.
upgrading my network
Which endpoint to add depends on what its user needs to do.
| endpoint | examples | ideal for |
|---|---|---|
| computers | desktops, laptops, servers | tasks that need high interactivity, multitasking or processing power: content creation, software development, data analysis |
| IoT devices | thermostats, appliances, smart TVs, security systems, fitness trackers, medical monitors | automating tasks, monitoring environments, streaming media, remote control |
| mobile devices | smartphones, tablets, handheld gaming devices | flexible, on-the-go communication, quick data entry, mobile access to applications |
| specialized | printers and scanners, VoIP phones, point-of-sale (POS) terminals | documents; voice calls over the internet; transactions and inventory |
Internet of Things (IoT) devices are everyday items able to connect to other devices and the internet. Mobile devices are small, lightweight endpoints that connect wirelessly.
A video editor needs a computer; a store needs a POS terminal; a home that wants its heating on before anyone arrives needs a smart thermostat.
cable or wireless
Choosing a medium for a LAN means weighing device location, performance needs, susceptibility to interference, ease of installation and cost. The best choice balances performance, reliability and budget for the users.
| medium | cost | installation | speed | interference |
|---|---|---|---|---|
| Twisted pair (UTP or STP) | affordable | easy | up to 10 Gbps, up to 100 m | STP for high EMI; UTP for most installations |
| Coaxial | moderate | more rigid, harder than twisted pair | up to 1 Gbps; significant loss over 450 m | better shielding than twisted pair |
| Fiber optic | most expensive | fragile, special tools | 1–100 Gbps, up to 100 km | immune: light, not electrical pulses |
| Wireless (Wi-Fi) | affordable | easy, no cables | 500 Mbps–1 Gbps, varies with signal | most susceptible: walls, devices, congestion |
Twisted pair comes unshielded (UTP) or shielded (STP), which protects better against interference. It suits devices that need a faster, more stable connection than wireless: desktops, printers, NAS. Use STP where EMI (electromagnetic interference) is high, near HVAC equipment, fluorescent lighting or other electronics that emit radio signals.
Wireless suits devices that need mobility: phones, tablets, laptops, IoT. Coaxial fits networks that already use coaxial connectors, often in older buildings with cable internet. Fiber optic is for ultra-high bandwidth, complete EMI resistance or maximum stability; it is rarely used inside SOHO LANs.
In the figure each medium carries a pulse past a fluorescent light: the noise is strongest on wireless and UTP, weaker on STP and coax, and none on fiber.
DHCP, or an address that never changes
A dynamic IP address is assigned automatically and may change when the device reconnects or after a set period. A static IP address stays the same.
Dynamic addressing suits networks that need flexibility and efficient use of address space. DHCP automates assignment and configuration, giving addresses only to devices currently connected.
In the figure (hypothetical) the router hands out addresses from 192.168.1.100 to 192.168.1.199, 100 in all. A laptop, phones and a guest tablet come and go; each takes the lowest free address, and an address freed by one is reused by the next.
Static addressing is for devices that need a consistent address: servers, printers, remote services, security systems. Here the printer, security camera and NAS keep 20, 30 and 50, outside the pool. Dynamic suits mobile or temporary devices: tablets, laptops, guests.
Dynamic addressing is easier with little administrative support, so it is common in homes and small offices. Static addressing gives precise control, but needs manual setup and careful tracking to avoid IP address conflicts.
advanced features on my network
AI tools can suggest a SOHO network design, configuration or fix. A useful prompt gives details: the number of users, the types of devices, how the internet is used and the budget. Vague prompts can lead to impractical or overly complex advice.
AI suggestions may include unrealistic devices, overcomplicated setups and incorrect IP addressing schemes. Some are technically accurate but wrong for the network at hand.
Judge a suggestion by whether it meets the needs for performance, ease of setup, wireless coverage and cost-effectiveness. It should keep the simplicity a small office or home expects, and be realistic, secure and user-friendly.
In the figure (hypothetical) the prompt is 4 users, 4 laptops, a printer, 2 phones, a $400 budget. The AI suggests a 48-port managed switch, three ceiling access points, a firewall appliance and a rack server: $3,750, on 172.32.0.0, which is not a private range. A wireless gateway and a small switch, $210, meet every need.
local files, games and streaming
A media server is a device on the network that streams content to other devices, stores shared files and hosts multiplayer games. A local one gives more control, faster access and more privacy than a cloud or subscription service, and often costs less.
To set one up:
In the figure (hypothetical) a wired NAS at 192.168.1.50 shares 4 folders; the TV streams a film, the laptop opens a document and the console joins a game.
identifying security needs
A SOHO network without proper security controls risks unauthorized access, exposure or loss of sensitive data, and disruption of services such as internet, file sharing or printing. Weak passwords, default configurations and unmonitored devices raise the risk.
Routers and access points left on default usernames, passwords or settings, widely known or easily guessed, let an adversary in to change the network’s configuration or disable its security controls.
Devices that are not updated or secured, with outdated firmware, or untrusted devices with known vulnerabilities, can lead to data loss, device compromise or malware infection.
Once in, an adversary can move laterally to other systems and devices, often through insecure IoT devices or unknown guest devices with weak credentials or unmonitored access. In the figure an adversary gets into a smart camera still on its default password admin/admin, and from there moves to the laptop and then the NAS: one network, nothing in the way.
mitigating the vulnerabilities
To choose a control, first identify the vulnerability and how it could affect data, devices or services. SOHO vulnerabilities often come from misconfigurations, outdated software or insecure devices.
| vulnerability | controls |
|---|---|
| default configurations on routers and access points | change the default login credentials; enable WPA2 or WPA3 wireless encryption; disable WPS |
| firmware | install firmware updates regularly, for the security patches; enable automatic updates if available |
| IoT devices | change their default credentials; patch firmware and software regularly; isolate them on a network segment separate from critical devices and data, to limit lateral movement |
| unknown devices | a guest network separate from the main one: visitors reach the internet, internal devices and data stay isolated |
In the figure the same attack meets the controls: the IoT devices sit on their own segment and guests on their own network. Even from the camera, the adversary gets no further than the segment’s edge; with the camera’s default password changed, it would not have got in at all.
applying defense
Firewalls and network segmentation are technical controls that limit unauthorized access by restricting how traffic flows through a network.
Together they create zones of security, such as guest traffic kept apart from internal systems, which helps stop malware or adversaries moving laterally.
A firewall limits the traffic allowed into or out of a network by applying rules that evaluate IP addresses, ports and protocols. Less traffic allowed means less risk of data exposure. A firewall checks a packet against its rules from the top, and the first rule that matches decides; a last rule denies everything else. The figure’s rules (hypothetical):
| # | action | direction | protocol | port | for |
|---|---|---|---|---|---|
| 1 | allow | out | TCP | 443 | web browsing (HTTPS) |
| 2 | allow | out | UDP | 53 | name lookups (DNS) |
| 3 | deny | in | TCP | 3389 | remote desktop from outside |
| 4 | deny | any | any | any | everything else |
Segmentation divides a network into smaller segments: it isolates sensitive or high-risk devices and data, reduces unnecessary traffic and creates distinct security zones.
segments for performance and security
Segmentation groups devices into zones by shared role, security risk level or performance needs. Weigh each device’s function, its security risks and its performance requirements.
| group by | to | for example |
|---|---|---|
| role | simplify troubleshooting and expansion | security camera systems; media devices (game consoles, streaming servers); family devices (laptops, tablets, phones) |
| risk level | isolate vulnerable or untrusted devices, so threats cannot spread | office computers and servers that need high security; IoT devices (thermostats, speakers, smart TVs), which are often vulnerable; guest devices on a separate wireless segment |
| performance needs | prioritize high-bandwidth or low-latency traffic | consoles and computers for online play; media servers for streaming; devices for remote video editing or rendering |
| type of user | reduce unnecessary traffic and simplify access control | office departments such as sales and accounting; teachers, students and technicians |
The figure groups the same 13 devices (hypothetical) by role, then by risk level, then by performance needs. A device can belong to different zones under different plans: the smart TV is a media device by role, an often-vulnerable IoT device by risk, and a streaming device by performance.