Cisco Certified Network Associate (CCNA) 200-301 · Free study guide

Objective 1.2 — Compare network topology architectures

A topology is a set of traffic paths and failure domains, not just a drawing. Compare architectures by scale, traffic direction, policy placement, operational complexity, and the consequence of losing a device or link.

Hierarchical campus designs

A two-tier campus combines core and distribution functions into one layer above access switches. It fits sites where scale and fault boundaries do not require a separate core. A three-tier design separates access, distribution, and core. Distribution aggregates access and applies routing or policy; the core provides fast transport between distribution blocks. The extra layer can improve scale and modularity, but it adds devices and paths that must be operated.

Redundancy is not merely drawing two upstream lines. Layer 2 loop prevention, Layer 3 routing, gateway redundancy, and failure convergence must agree.

Spine-leaf fabrics

In a spine-leaf design, each leaf connects to each spine, and endpoints attach to leaves. Traffic between leaves has a predictable leaf-spine-leaf path. The design supports east-west data-center traffic and equal-cost routed paths. Spines do not normally connect directly to endpoints, and adding a spine expands fabric path capacity without changing the leaf attachment model.

Do not assume spine-leaf automatically provides segmentation or policy. The underlay supplies transport; overlay and policy mechanisms provide logical services when the architecture uses them.

WAN, SOHO, cloud, and on-premises

A WAN joins separated sites or provider domains and introduces carrier circuits, Internet transports, latency, and shared responsibility. A SOHO design is small, but it still needs secure wireless, addressing, updates, backups, and recovery. Its functions may be consolidated into fewer devices.

On-premises infrastructure is operated in facilities the organization controls. Cloud infrastructure moves some physical ownership to a provider and exposes services through provider constructs. A hybrid design connects the two. Cloud does not eliminate routing, identity, segmentation, or monitoring; it changes interfaces and responsibility boundaries.

Worked scenario

A small headquarters with four access closets can use a redundant two-tier pair if its scale and port needs fit. A multi-building campus may benefit from separate distribution blocks and a core. A data center with heavy server-to-server flows may use leaf-spine. Remote branches traverse a WAN, while a small home office may combine routing, switching, firewall, and AP roles. The correct answer follows the requirements, not the newest terminology.

Verification questions

For any diagram, ask: where are the Layer 2 boundaries, which devices route, where is policy enforced, what path does east-west and north-south traffic take, and what fails when one link or node disappears? Then identify which routing or loop-prevention mechanism supports the drawn redundancy.

Common traps

• Calling every collapsed design three-tier because it has three device types.
• Treating two parallel links as resilient without a convergence mechanism.
• Assuming the cloud provider owns customer addressing and policy decisions.
• Using spine-leaf as a synonym for any redundant topology.

Readiness checklist

• I can distinguish two-tier, three-tier, and spine-leaf traffic paths.
• I can describe WAN and SOHO operational constraints.
• I can separate placement from responsibility in cloud and on-premises designs.
• I can name the failure domain introduced by each shared component.