Enterprise network design: Layers and redundancy for growth and resilience
Enterprise network design is not based on headcount alone. Communication needs, capacity, acceptable downtime, and access policies determine device roles and network layers. Apply Cisco's design guidance to networks ranging from a single router to multiple wiring closets and buildings. Plan addresses, link capacity, failure domains, and L2/L3 boundaries, then consider the layers and redundancy you need.
The introduction separates the functions combined in a single router. Then follow four sections to explore specific design criteria and configurations.
After the core design sequence, related articles cover security, validation, incident response, and connections to external networks.
Before reading a design, write down which business activities need to communicate, between which locations, and how much interruption is acceptable. Use those questions to divide equipment responsibilities in the articles on network layers, then compare configurations from small sites to campuses. Rather than choosing a layer structure solely from headcount or device count, check whether the communication scope and operating assumptions of each example match your environment.
When reading about link capacity and failure domains, draw both the normal flow and the paths left when one link or device is unavailable. Connect the L2/L3 boundary and redundancy articles to ask how far the effects of a change might spread. An address range checked with a calculator is only one part of a design; assessing the configuration continues with the checks discussed in validation and incident response. After the core design sequence, add external connectivity and security as separate requirements.
Four sections in this chapter
Explore four sections covering network roles, small-site layouts, capacity and failure domains, and network layers and redundancy.
How do large enterprise networks split a home router's roles?
How enterprise networks divide switching, routing, NAT, firewall, and DHCP roles by failure impact, change scope, ownership, and redundancy needs.
Start with business needs, then define roles and layers
First, see why you would separate the functions combined in a router. Identify who uses each service and how much downtime they can tolerate. Then compare those needs with Cisco's reference designs and the roles of Access, Distribution, and Core.
Choose equipment and plan addresses for a small site
Compare when one router is enough with when you need separate switches and APs. Check ports, PoE, wireless requirements, and the gateway, and plan addresses that will not overlap when you connect to branches or the cloud.
Set link capacity, failure domains, and L2/L3 boundaries
Calculate the capacity needed both during normal operation and on the paths left after a failure. Distinguish the effects of shared power and cable pathways from equipment failures. Then decide how far to extend each VLAN and where to separate networks with routing.
Connect wiring closets and buildings with redundancy
Start with a two-tier design that aggregates access switches. Then examine failure detection and path switching for switches, gateways, and routers. Finally, decide when you need a separate Core to connect multiple distribution blocks.
Background reading
These related articles are outside the core reading sequence.
- Whose MAC address does my PC use to send data? — Review IP and MAC roles, and the different decisions made by a switch and a router
- How do you separate devices on the same switch so they cannot communicate directly? — Review the communication domains of VLANs and trunks
- How does a router choose where to send a packet? — Review how a router selects a route
Explore further after the core articles
These related articles are outside the core reading sequence.
- How do enterprise networks connect to branches, clouds and the internet? — Connect the internal network you designed to branches, the cloud, and the internet
- How do you add security to a network design? — Put the access policies defined at the start into practice with security controls
- How do you test whether your network design can withstand failures? — Test the planned paths and recovery targets
- What should you check and restore first when a network fails? — Respond based on business impact and evidence