8 Integrated Network Design and Practice
Learn how physical links, switching, VLANs, IP addressing, routing, and troubleshooting fit together in a small campus or office network.
How network components fit together
A working network depends on several layers fitting together: physical links carry Ethernet frames, switches place devices in the appropriate local network, and IP addressing and routing deliver packets to destinations. A is a common design for a small campus or office: access switches connect computers, printers, servers, and wireless access points; those switches connect to a core or Layer 3 ; and a or connects toward the Internet.
Switching, routing, and access
A connects devices within a local area network (LAN) and forwards Ethernet frames using MAC addresses. A connects different IP networks and forwards packets using IP addresses; a Layer 3 can also route between local networks. A applies traffic rules, often at the LAN–Internet boundary.
A connects Wi-Fi clients to the wired LAN. It is not necessarily a : many APs bridge wireless clients onto a network connected through a .
Ethernet links and boundaries
Ethernet belongs to the IEEE 802.3 family and operates over media such as twisted-pair copper and fiber. Choose media and transceivers to match the required speed, distance, and equipment. Properly terminate and test cables, then check link indicators and negotiated speed before investigating IP settings. Fiber is often suitable for longer links or links between buildings; copper Ethernet is common for endpoint connections.
A divides a switched network into a separate Layer 2 broadcast domain. Devices in different VLANs generally need a or Layer 3 to communicate. separation can organize users, servers, or IoT devices, but does not by itself provide a complete security policy.
IPv4 prefixes and subnet sizing
An IPv4 address has bits. places the network prefix length after a slash. For example, the prefix in 192.168.40.0/26 uses the first bits to identify the prefix, leaving bits for addresses within the block. The block therefore has total addresses. In a conventional LAN subnet, the first address is the network address and the last is the broadcast address, leaving usable host addresses.
Private IPv4 ranges include 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16. Address pools must not overlap with statically assigned addresses.
addressing
addresses have bits, and a typical LAN is assigned a /64. An interface may also have a , which is used only on its local link. does not use an IPv4-style broadcast address.
The prefix 2001:db8::/32 is reserved for documentation and examples, not production addressing. In a real network, use prefixes assigned or generated for that deployment.
A two- addressing plan
The example plan assigns a staff and an IoT separate IPv4 subnets and documentation prefixes:
Staff:
10;192.168.40.0/26; example gateway192.168.40.1; documentation prefix2001:db8:40:10::/64.IoT:
20;192.168.40.64/27; example gateway192.168.40.65; documentation prefix2001:db8:40:20::/64.
For the staff subnet, hosts can use 192.168.40.1 through 192.168.40.62; 192.168.40.0 is the network address and 192.168.40.63 is the broadcast address. The /26 block has usable host addresses.
For the IoT subnet, hosts can use 192.168.40.65 through 192.168.40.94; 192.168.40.64 is the network address and 192.168.40.95 is the broadcast address. The /27 block has usable host addresses. These subnet sizes are sufficient for needs of staff addresses and IoT addresses, respectively.
The example prefixes are documentation space, not production assignments. Keep boundaries consistent for IPv4 and , and configure and test routing and filtering for both protocols.
Troubleshooting from the link upward
When a device cannot reach a service, check the network in order rather than changing settings at random:
Physical link: Check power, cable seating, port status, Wi-Fi association, and link speed. Try a known-good cable or port.
Ethernet and : Confirm that the learned the device's MAC address on the expected port and that the port is assigned to the intended . A working link does not prove the is correct.
IP configuration: Check the address, prefix or subnet mask, and . Confirm that the host is in the subnet expected for its and that DHCP supplied valid settings.
Local reachability: Test the . If it does not respond, focus on host configuration, the , the -to- link, or the gateway interface.
Routing and policy: If the gateway responds but a remote subnet does not, check routes, , and rules. Repeat these checks for IPv4 and ; success on one protocol does not prove the other works.
Name and application: If an IP destination works but its hostname does not, investigate . If both work but the application fails, check the service, port, and application-level policy.
Useful host-side checks include viewing interface configuration, checking the routing table, testing the gateway and destination with ping, and tracing the path with a traceroute utility. Interpret results carefully: a device or may block diagnostic probes even when other traffic works.
Diagnosing cross-network and protocol failures
If a laptop has a valid address and can reach its gateway but cannot reach a server in another , check , the destination route, and rules between the VLANs. Also confirm the server's gateway and address plan.
If a service works over IPv4 but not , verify the host's address and prefix, its default route, the 's interface and route, and -specific policy. If a hostname is used, check records as well; also verify that the service listens on . Working IPv4 does not prove is configured correctly.