Planning the Network Infrastructure
What the plan must contain
Infrastructure planning starts with requirements, not products. A defensible plan documents the business drivers, the current-state inventory (subnets, link speeds, device ages, licence entitlements), the target topology, and the constraints (budget, floor plans, cabling routes, power). Candidates should be fluent in the vocabulary of segmentation: broadcast domains, VLANs, subnets, and the routing boundary between them. Know how to size a subnet from a host count, how to reserve ranges for servers, printers, voice, guests, and management, and why a management VLAN is kept separate from user traffic.
Capacity and availability
Be able to estimate growth and translate it into address space and bandwidth. Understand redundancy choices - redundant links, redundant power, dual-homed servers - and the difference between redundancy and backup. Site planning concepts (WAN link sizing, local internet breakout, DHCP relay placement) appear frequently.
Traps
Questions love to blur planning with implementation. If the stem asks what should be documented or decided before deployment, an answer that describes a configuration step is usually wrong. Another classic trap: assuming one subnet equals one physical location, or forgetting that a routed network needs a default gateway per subnet. Watch for subnet-sizing items where the "obvious" mask wastes half the range; always confirm usable hosts (2^n − 2).
Rule of thumb
Plan address space for at least double current need, and never allocate the smallest subnet that fits today. For example, a site needing 50 hosts should get a /26 (62 usable) rather than a /27 (30 usable) only if growth is credible; if 50 is a hard ceiling, /26 is still the safer default because it leaves room for a switch, printer, and access point you forgot.
Also know
Naming conventions, IP address management ownership, documentation of cable runs and patch panels, and the change-control process that turns a plan into an approved baseline.