The three-tier design is where enterprise hierarchy begins, and you already know it from your CCNA.
This review covers it first, then the two-tier and the spine-and-leaf, the three designs ENCOR expects you to recognize on sight.The Three Layers
The hierarchy splits the network into three layers:
Access is where users and devices plug in, mainly Layer 2.
Distribution aggregates the access switches and applies policy.
Core is the high-speed backbone between distribution blocks.

Figure 1 – The three-tier model: Access, Distribution, and Core
One access switch with its distribution forms a repeatable unit, called a switch block.
You scale the network by adding blocks, not by redesigning it.Access Layer
The Access Layer is the edge where users connect.
Laptops, IP phones, printers, and access points all plug in here.
Figure 2 – The Access Layer connects end-user devices
It runs mostly at Layer 2, forwarding by MAC address.
It also delivers edge services and security: VLANs, STP, and Port Security.Answer the question below
Which layer is the network's edge, where users plug in?
Distribution Layer
The Distribution Layer aggregates every access switch in a block.
It works at Layer 3, routing between VLANs and toward the core.
Figure 3 – The Distribution Layer aggregates and routes
This is where the Layer 2 to Layer 3 boundary sits.
Policy lives here too: ACLs, QoS, and route summarization toward the core.Core Layer
The Core Layer is the backbone.
Its only job is to move traffic between distribution blocks as fast as possible.
Figure 4 – The Core Layer is the high-speed backbone
Keep the core simple.
You avoid ACLs, NAT, and filtering here, because every extra operation adds latency to traffic passing through.In a multi-building campus, each building keeps its own access and distribution.

Figure 5 – Multiple buildings interconnected through a centralized core
The core sits in the data center and links every block through redundant high-speed paths, usually fiber with EtherChannel.
Here are the three layers side by side:
Layer
Main Function
OSI Layer
Key Technologies
Access
Connects end devices (PCs, IP phones, printers, APs)
Layer 2
VLANs, PoE, STP, Port Security, DHCP Snooping
Distribution
Aggregates access, routes between VLANs, applies policy
Layer 3
L3 switches, Inter-VLAN routing, ACLs, QoS
Core
High-speed backbone between distribution blocks
Layer 3
Fiber, 10/40/100/400 Gbps, EtherChannel
Table 1 – Three-tier roles and technologies
Answer the question below
Which layer should stay free of ACLs and filtering?
Now shrink the three-tier.
In smaller sites, a dedicated core is wasted hardware, so you merge two layers into one.Merging Distribution and Core
The two-tier design, also called the collapsed core, combines the distribution and core onto one switch pair.

Figure 6 – Three-tier vs collapsed core: two layers merged into one
The access layer does not change.
Users still connect to Layer 2 access switches, exactly as before.
Figure 7 – A typical collapsed core network
One pair of multilayer switches now carries the whole upper network.
Inside the Collapsed Core
That collapsed-core pair takes on several jobs at once:
Aggregates traffic from all access switches.
Routes between VLANs.
Provides the exit toward the WAN, Internet, or other networks.

Figure 8 – The collapsed core aggregates and routes all traffic
To stay highly available, it runs the redundancy you already know.

Figure 9 – SVI gateways, HSRP redundancy, and STP loop prevention
Expect SVIs as VLAN gateways, HSRP or VRRP for gateway redundancy, and STP between the two tiers.
Answer the question below
How many switch layers does the collapsed core design have?
When to Use It
Fewer layers mean a simpler, cheaper network, with a trade-off:
Advantages
Limitations
Simplified design: easier to deploy and manage.
Limited scalability: not ideal for very large sites.
Cost-effective: no dedicated core switches.
Possible bottlenecks if too many access switches are added.
Table 2 – Collapsed core: advantages and limitations
Use two tiers for small and medium campuses and branch sites.
Keep three tiers when the network must scale to many buildings or thousands of users.Answer the question below
Which two layers merge in a collapsed core?
The last design breaks from the hierarchy entirely.
You saw spine-and-leaf in your CCNA, built for the data center, where traffic moves differently.Why the Data Center Changed
Early data centers carried mostly North-South traffic: clients in, servers out.
The three-tier model handled that fine.Then virtualization flipped the pattern.
Most traffic became East-West, server to server, inside the data center.
Figure 10 – East-West traffic exposes the three-tier limits
Here the three-tier hits a wall.
STP blocks the redundant links, so bandwidth is wasted and every server-to-server flow crosses extra layers, adding latency.Answer the question below
Which traffic pattern dominates inside the data center?
The Spine and Leaf Fabric
Engineers asked a simple question: what if every link were Layer 3?
No Spanning Tree, no blocked links, every path active.That idea became the spine-and-leaf fabric.
Leaf switches connect your servers and storage, and spine switches form the backbone.One wiring rule governs the fabric:
Every leaf connects to every spine, a full mesh.
Spines never connect to spines, and leaves never connect to leaves.

Figure 11 – Spine-and-leaf: every leaf meshed to every spine, Layer 3 throughout
Because every link is routed, every link is active.
There is no Spanning Tree and nothing is blocked.Answer the question below
Do spine switches ever connect directly to each other?
Always the Same Distance
This layout makes performance predictable.
Any server reaches any other in exactly two switch hops: leaf to spine to leaf.
Figure 12 – Two hops between any two servers
Need more bandwidth? Add another spine.
Every leaf gains a new active path, and capacity grows without a redesign.Spine-and-leaf gives your data center a fast, predictable wiring pattern.
It is often called a fabric, and that word deserves a closer look.Answer the question below
How many switch hops separate any two servers in spine-and-leaf?
Answer the question below
In a three-tier campus, how many hops link access switches in two different buildings?