LAN Architecture
Common Network Topologies
- Concept: The physical or logical layout of how network devices are interconnected.
- Details:
- Star Topology: Multiple devices connect to a single central device (e.g., PCs connecting to a central switch).
- Full Mesh Topology: Every device is connected to every other device (e.g., 6 routers fully interconnected). Provides maximum redundancy but high cost.
- Partial Mesh Topology: Some devices are connected to each other, but not all (e.g., top switches connected to all, bottom switches not directly connected to each other).
- Hybrid Topology: A combination of two or more different topologies.
graph LR
subgraph Star
C((Switch))
A[PC 1] --- C
B[PC 2] --- C
D[PC 3] --- C
end
subgraph Full Mesh
R1[R1] --- R2[R2]
R1 --- R3[R3]
R2 --- R3
end
subgraph Partial Mesh
SW1[SW1] --- SW2[SW2]
SW1 --- SW3[SW3]
SW2 --- SW3
SW2 --- SW4[SW4]
SW3 --- SW4
end
Things to note: In network diagrams, devices might not be drawn in a literal star shape, but if many devices connect to one central device, it is logically a star topology.
Traditional Campus LAN Design
- Concept: Hierarchical network design used in enterprise networks (buildings or nearby buildings).
- Details: Designed to be scalable, manageable, and resilient.
Two-Tier LAN Design (Collapsed Core)
- Concept: Consists of two layers: Access Layer and Distribution Layer. Omits the Core layer (or combines Core and Distribution into one layer).
- Details:
- Access Layer:
- End hosts (PCs, printers, cameras, WAPs) connect here.
- High port density.
- Quality of Service (QoS) marking is typically done here (mark traffic as early as possible).
- Security services: Port security, Dynamic ARP Inspection (DAI), DHCP snooping.
- Power over Ethernet (PoE) enabled switchports for WAPs and IP phones.
- Distribution Layer:
- Aggregates connections from Access layer switches (typically a redundant pair).
- Border between Layer 2 (L2) and Layer 3 (L3).
- Runs both L3 (OSPF) and L2 (Spanning Tree Protocol - STP).
- Connections from Access to Distribution are typically L2 (STP blocks links to prevent loops).
- End hosts use Switch Virtual Interfaces (SVIs) on Distribution switches as default gateways.
- Uses First Hop Redundancy Protocols (FHRPs) like HSRP or VRRP for redundant virtual IPs.
- Connects to services (Internet, WAN) and other parts of the LAN.
- Also called the Aggregation Layer or Core-Distribution Layer.
- Access Layer:
Three-Tier LAN Design
- Concept: Adds a Core Layer to the Two-Tier design. Recommended by Cisco if there are more than three distribution layers in a single location.
- Details:
- Core Layer:
- Connects separate Distribution layers together in large LANs.
- Focus: Speed and “fast transport”.
- Avoid CPU-intensive operations (security features, QoS marking/classification). Just fast packet forwarding.
- All connections are Layer 3.
- No Spanning Tree Protocol (STP) at the Core Layer.
- High redundancy for devices and connections (backbone of the LAN).
- Core Layer:
graph TD
subgraph Core Layer
C1[Core Switch 1] --- C2[Core Switch 2]
end
subgraph Distribution Layer
D1[Dist Switch 1] --- D2[Dist Switch 2]
end
subgraph Access Layer
A1[Access Switch 1]
A2[Access Switch 2]
end
C1 --- D1
C1 --- D2
C2 --- D1
C2 --- D2
D1 --- A1
D1 --- A2
D2 --- A1
D2 --- A2
Things to note: In a 2-tier design, the distribution layer must fulfill the role of the core layer, which is why it is sometimes called the Core-Distribution layer.
Spine-Leaf Architecture (Data Center)
- Concept: A modern 2-tier architecture designed for data centers to handle high East-West traffic (server-to-server traffic) caused by virtualization and distributed applications. Also known as Clos architecture.
- Details:
- Spine Switches:
- Form the backbone of the fabric.
- Connect to every Leaf switch.
- Do not connect to other Spine switches.
- Leaf Switches:
- Connect to end hosts (servers) - acts like the Access layer.
- Connect to every Spine switch.
- Do not connect to other Leaf switches.
- Key Characteristics:
- Consistent Latency: Every server is separated by the exact same number of hops (e.g., Leaf -> Spine -> Leaf), providing predictable latency for East-West traffic.
- Load Balancing: Traffic paths are randomly chosen to balance load among Spine switches.
- Scalability: Easily scaled by adding a new Leaf switch and connecting it to all existing Spine switches.
- Spine Switches:
graph TD
subgraph Spine Layer
S1[Spine 1]
S2[Spine 2]
S3[Spine 3]
end
subgraph Leaf Layer
L1[Leaf 1]
L2[Leaf 2]
L3[Leaf 3]
end
subgraph Servers
Server1[Server 1]
Server2[Server 2]
Server3[Server 3]
end
L1 --- S1
L1 --- S2
L1 --- S3
L2 --- S1
L2 --- S2
L2 --- S3
L3 --- S1
L3 --- S2
L3 --- S3
Server1 --- L1
Server2 --- L2
Server3 --- L3
Things to note: Traditional 3-tier designs caused bottlenecks and variable latency for East-West traffic because traffic had to traverse up to the core and back down. Spine-leaf solves this by keeping the hop count consistent.
SOHO (Small Office / Home Office) Networks
- Concept: Small networks (e.g., home networks, small businesses) with simple needs and few devices.
- Details:
- All networking functions are consolidated into a single multipurpose device, often called a Home Router or Wireless Router.
- Combined Functions:
- Router: Connects the internal network to the Internet (ISP).
- Switch: Provides a few physical switchports for wired devices.
- Firewall: Typically blocks inbound unsolicited connections while allowing outbound internal connections.
- Wireless Access Point (WAP): Provides Wi-Fi connectivity.
- Modem: Sometimes integrated for cable/DSL Internet (though often a separate device).
Things to note: Enterprise networks use dedicated devices for each function, but SOHO networks combine them into one small, inexpensive device.
Comparison of LAN Architectures
| Feature | Traditional 2-Tier | Traditional 3-Tier | Spine-Leaf |
|---|---|---|---|
| Layers | Access, Distribution | Access, Distribution, Core | Spine, Leaf |
| Primary Use Case | Small/Medium Enterprise | Large Enterprise Campus | Modern Data Centers |
| L2/L3 Boundary | Distribution Layer | Distribution Layer | Leaf Switches (typically) |
| STP Usage | Yes (Access/Distribution) | Yes (Access/Distribution) | No (Uses L3 routing) |
| Traffic Focus | North-South | North-South | East-West |
| Scalability | Moderate | High | Extremely High |
Important Concepts & Misconceptions
- Misconception: The Core layer is just a “faster” Distribution layer.
- Correction: The Core layer is strictly Layer 3 and must not perform packet manipulation (like ACLs, QoS marking, or STP). Its sole focus is fast transport.
- Misconception: Spine-Leaf is a 3-tier architecture.
- Correction: Spine-Leaf is a strictly 2-tier architecture (Spine and Leaf).
- Rule of Thumb: Cisco recommends adding a Core layer if you have more than 3 Distribution blocks in a single location.
- FHRP & STP Interaction: In a 2-tier/3-tier design, STP runs at the Access/Distribution boundary (L2), while FHRPs (HSRP/VRRP) provide default gateway redundancy at the Distribution layer (L3 SVIs).
Commands Summary
Note: This lesson focuses on network design theory and architecture. There are no specific IOS CLI commands introduced in this transcript.