Designing a scalable campus network

Most campus networks that "don't scale" weren't badly built — they were built for the size the campus was on day one, with no seams for the size it became. Scalable design isn't about over-building; it's about leaving the right seams in the right places.

The hierarchical model still holds up

Access, distribution, core — the classic three-tier model earns its longevity because each layer has one job. Access switches terminate end devices and enforce port-level policy. Distribution aggregates access switches, routes between VLANs, and is where policy boundaries usually live. Core exists purely to move traffic between distribution blocks as fast as possible — no policy, no VLANs terminating there, just speed.

The mistake that breaks scalability is collapsing layers too early to save on hardware, then discovering the "temporary" collapsed design is now load-bearing for three buildings you didn't originally plan for.

Design in blocks, not one flat topology

Instead of one core connecting to every access switch on campus, group access switches into a distribution block per building or floor, and connect blocks to the core. Adding a new building becomes "add a block to the core" instead of "redesign the whole campus." This is the single highest-leverage decision in campus design — it turns growth into an additive operation instead of a disruptive one.

Leave headroom in the addressing plan

A /24 per VLAN feels generous until a building triples its device count with IoT sensors and access points. Summarize at the distribution layer so the core's routing table stays small, but leave enough address space per block that growth doesn't force a re-addressing project. Renumbering a live building is one of the most disruptive projects a network team can be handed — a design that avoids it earns its complexity budget many times over.