Network topology architectures
Various network topology architectures
Networks come in many designs and connectivity models. The right design depends on two main things:
- Scale (small, medium, or large)
- Purpose (home office, enterprise/campus, data center, or service provider) As a network engineer (and a CCNA candidate), you’ll want to recognize these common topology architectures and understand why each one is used.
Small office/home office (SOHO) networks
Most people are already familiar with the small office/home office (SOHO) network type. If you have internet at home, you already have a SOHO-style network. A typical SOHO network includes:
- Endpoints (PCs, laptops, phones, TVs)
- A wired router with a built-in switch, or a wireless router/wireless access point
- A cable modem or fiber optic modem (depending on your internet service) It’s also common for the modem to include routing and switching features (an “all-in-one” device).
If a home office has more devices than the built-in switch can support, you’ll often see an additional switch added for more wired ports. For wireless coverage, you may also see wireless extenders or additional wireless access points placed around the home to extend coverage. This same general design is also common in small branch offices, where keeping costs low matters. These offices often use lower-end switches throughout the space that connect back to a single router, which then connects to broadband internet (cable or consumer-grade fiber). To support more users and applications, these sites often pay for:
- More bandwidth
- Higher speeds to support more throughput (the amount/rate of traffic that can pass through a network over time)
Enterprise (Campus) networks
The next topology architecture is commonly used in medium to large businesses (often called enterprises). These organizations may have multiple buildings in close geographic proximity, with their office networks interconnected. That’s why these are called campus networks - the design is similar to a college campus. As enterprise networks grow, they’re often built in tiers (also called layers). Each layer has devices with specific roles, which helps the network stay efficient, resilient, and easier to manage. Enterprise campus networks typically require more resiliency and redundancy than SOHO networks. Downtime is expensive, so these networks usually use enterprise-grade equipment designed for heavy usage and high reliability. You’ll also see redundant devices and links to support business continuity if a device or connection fails.
Two-tier (Collapsed Core) network architecture
For many small to medium enterprise networks, a two-tier architecture is enough. This topology has two layers:
- Access layer
- Distribution layer
Access Layer
The Access Layer is where endpoints connect to the network. This is where you’ll find switches that connect:
- Endpoints
- Servers
- Wireless access points Wireless endpoints connect to the wireless access points, and those access points are hardwired into the Access Layer switches.
Distribution Layer
The Distribution Layer sits above the Access Layer. It typically contains a pair of switches that all Access Layer switches connect into. Key ideas in this layer:
- There are usually two distribution switches for redundancy.
- Access switches often have redundant uplinks (more than one) to the distribution switches.
- If one uplink fails or is disconnected, traffic can continue over the remaining uplink.
Three-tier network architecture
A three-tier topology keeps the same Access and Distribution Layers, but adds a third layer: the Core Layer.
- The Distribution Layer uplinks to a pair of Layer 3 switches in the Core Layer.
- This design is typically used in large enterprise campus networks. It’s especially useful when you have three or more buildings or multiple geographically dispersed sites. Without a Core Layer, you may end up connecting distribution blocks to each other in a full mesh, which doesn’t scale well.
In a full-mesh distribution design, the number of links grows quickly as you add more buildings and distribution switches. That increases complexity and cost. It can be especially expensive if buildings must be connected using service-provider fiber. With a Core Layer:
- Each distribution block connects to the core switches.
- You avoid connecting every distribution block to every other distribution block.
- Each distribution block typically needs only two uplinks to the core, while still avoiding single points of failure.
The Core Layer is designed for high-speed, high-capacity routing between IP networks across the enterprise. Core switches are often located in the organization’s headquarters or in the data center.
In a three-tier architecture, the core switches connect to a redundant pair of routers, either owned/managed by the organization or provided by a service provider.
One of the main goals of the Core Layer is to provide high speed, uninterrupted routing. This is why you typically will not see a lot of security or traffic policies or restrictions at this layer.