Network Devices, Topologies & Virtualization

CCNA 200-301 · Network Fundamentals

By Abu Anns · Last updated 20 November 2026Exam version: 200-301 v1.1 (current until 2 Feb 2027)

Every network is built from a small set of device types, each with a job that the CCNA exam expects you to know precisely. The most common exam mistake in this topic is blurring the roles: a switch is not a small router, a router is not a firewall, and a controller does not forward user traffic. Learn each device by the question it answers — where does this frame go, where does this packet go, should this traffic be allowed at all — and the rest of the blueprint gets easier, because almost every later topic is one of these devices doing its job in more detail.

Network components and their roles

A router connects different IP networks and forwards packets between them based on the destination IP address. It is a Layer 3 device: it reads IP headers, consults its routing table, and makes a forwarding decision for each packet. Routers also mark the boundary of a broadcast domain — broadcasts do not cross a router — which is why every subnet needs a router (or a Layer 3 switch) as its gateway to the rest of the world.

A Layer 2 switch forwards Ethernet frames within a single network based on destination MAC addresses. It learns which MAC address lives on which port by watching source addresses on incoming frames, builds a MAC address table, and forwards each frame only out of the port where its destination lives. A Layer 3 switch does all of that and can also route between VLANs: it has a routing engine, so a single device can act as the switch for a floor of users and the router between their VLANs. In modern campus networks the Layer 3 switch has largely taken over the routing job inside a building, with dedicated routers handling the WAN edge.

A next-generation firewall (NGFW) inspects traffic and enforces security policy — it decides what is allowed to pass, based on far more than IP addresses and ports. Unlike a basic packet-filtering firewall, an NGFW understands applications, users, and content, and it can block a specific application or a malicious file inside otherwise permitted traffic. An intrusion prevention system (IPS) watches traffic for known attack signatures and suspicious behaviour and can block or drop malicious traffic in real time; an intrusion detection system only alerts, while an IPS acts. The two are often combined in one appliance, and the exam may ask which one merely detects and which one prevents.

An access point (AP) provides wireless connectivity, bridging wireless clients onto the wired network. In enterprise networks APs are usually lightweight: they handle the radio work while a wireless LAN controller (WLC) manages configuration, security, roaming, and RF settings centrally. A controller in the broader sense is the same idea applied to the whole network — a central brain that manages many devices, which is the foundation of the software-defined networking covered in the automation guide.

Endpoints are the devices users actually touch — laptops, phones, printers, IP phones, cameras — and servers are endpoints that provide services to others, such as file storage, DNS, or applications. Power over Ethernet (PoE) lets a switch deliver electrical power to an endpoint over the same Ethernet cable that carries data, which is how most enterprise access points and IP phones are powered without a separate power supply at each device. The original PoE standard (802.3af) delivers up to 15.4 watts per port, PoE+ (802.3at) up to 30 watts, and PoE++ (802.3bt) up to 60 or 100 watts for power-hungry devices such as pan-tilt-zoom cameras. If a device will not power on, checking the switch's PoE budget is a standard first troubleshooting step.

Topology architectures

Two-tier architecture (collapsed core) has an access layer connecting endpoints and a core layer that also does distribution work — two layers total. It is simple and cheap, and it suits small networks. Three-tier architecture adds a distribution layer between access and core: access switches connect endpoints, distribution switches aggregate access switches and enforce policy, and the core switches move traffic quickly between distribution blocks. Three-tier scales better and is the classic enterprise campus design. Spine-leaf architecture is the data-centre answer: every leaf switch connects to every spine switch, so any two devices are at most two hops apart, bandwidth scales by adding spines, and there is no blocking core. Traffic between servers in the same rack goes leaf-to-leaf through a spine; the design is built for the heavy east-west traffic between servers rather than north-south traffic to users.

WAN topologies describe how sites connect over distance — hub-and-spoke, partial mesh, and full mesh — and the trade-off is always cost against resilience: a full mesh survives any single link failure but needs a link between every pair of sites. SOHO (small office/home office) networks collapse everything into one or two devices: a single router with a built-in switch, wireless access point, and firewall. Finally, on-premises and cloud describes where services live: on-premises equipment sits in your own building and you manage it; cloud services run in a provider's data centre and you consume them over the internet. Most real networks are now a mix, and the exam wants you to know that cloud-hosted services change where traffic flows — user traffic that once stayed inside the building now goes out to the internet and back.

Virtualization fundamentals

Server virtualization lets one physical server run several virtual machines, each with its own operating system, sharing the hardware through a hypervisor. It matters for networking because a virtual machine still needs an IP address, a MAC address, and a path to the network — the physical server's network interface carries traffic for many virtual machines, and the virtual switch inside the hypervisor forwards frames between them. Containers are a lighter form of the same idea: they share the host's operating system kernel instead of each running a full OS, so they start faster and use fewer resources. A single host can run dozens of containers where it might run a handful of virtual machines.

A VRF (Virtual Routing and Forwarding) is virtualization applied to the routing table: one physical router maintains several separate routing tables, so traffic in one VRF is isolated from traffic in another even though it crosses the same hardware. VRFs are how service providers keep different customers' traffic apart on shared infrastructure, and how enterprises separate departments or security zones without buying a router per zone. The exam's key point is the concept — separate routing tables on one device — not the configuration.

One thread runs through all of this: modern networks push intelligence to the centre (controllers, virtualization, cloud) while keeping the forwarding fast at the edge. When you study the later guides, keep asking which device is doing the work and which layer it works at, and this foundation will keep paying off.

Advertisement

Exam traps to watch for

Practice questions

Test yourself on this guide. Answers are checked instantly, with a full explanation for each question.

Q1.A network engineer needs a device that forwards packets between two different IP subnets and stops broadcast traffic from spreading between them. Which device should be used?

Report an error

Q2.What is the primary difference between a Layer 2 switch and a Layer 3 switch?

Report an error

Q3.Which device inspects traffic at the application layer and can block a specific application even when the traffic uses an allowed port?

Report an error

Q4.A security appliance detects traffic matching a known attack signature and immediately drops the offending packets. Which type of system is this?

Report an error

Q5.A company is designing a data centre where servers exchange large amounts of traffic with each other, and any server must reach any other server in at most two hops. Which topology best fits this requirement?

Report an error

Q6.A new IP phone is connected to a switch port but will not power on. The port shows as up and the cable is known good. What is the most likely cause?

Report an error

Q7.What is the main difference between a virtual machine and a container?

Report an error

Q8.A service provider uses one physical router to carry traffic for two customers, and each customer's routes must be completely isolated from the other's. Which technology provides this?

Report an error

Q9.In a three-tier campus design, which layer aggregates access-layer switches and applies policy such as access control lists and route summarization?

Report an error