6. Ethernet and Switching

Introduction

Ethernet is the dominant technology used to connect devices inside most local networks. Switching is the method used to move Ethernet traffic efficiently between those devices. Together, Ethernet and switches form the foundation of almost every modern wired network, from a tiny home office to a large enterprise campus.

This chapter gives you a big picture of what Ethernet and switching are, how they fit together, and why they replaced older technologies like hubs. Later child chapters will dive into specific details such as MAC addresses, Ethernet frames, collision and broadcast domains, VLANs, and Spanning Tree Protocol. Here we stay at the conceptual level and focus on how all of these concepts relate to each other as a single system.

What Ethernet Is

Ethernet is a family of networking technologies that define how devices on a local network format data, address each other, and share the physical medium. It covers both how bits are carried on the wire and how devices on the same network segment decide who can talk and when.

Ethernet is standardized by the IEEE, mainly in the 802.3 set of standards. Over time, Ethernet has evolved from early speeds like 10 Mbps to 100 Mbps, 1 Gbps, 10 Gbps, and far beyond, but the basic ideas have remained consistent. It uses a frame based method to wrap data with headers and trailers that include addressing and error checking. Later chapters in this section will explain the exact structure of these Ethernet frames and the role of MAC addresses.

Ethernet usually operates on twisted pair copper cables or fiber optic cables, although it can also be used over other media. Inside buildings, Ethernet is what carries your data from computers, IP phones, printers, access points, and servers to the central switching infrastructure.

Shared Media vs Switched Ethernet

Early Ethernet networks used shared media. Many devices were attached to a single cable or connected using a hub, so all of them shared the same bandwidth. Only one device could effectively talk at a time. If two tried to send data at once, their signals interfered and caused a collision.

This original approach relied on a method called CSMA/CD to detect collisions and recover, which limited performance and made networks less predictable as more devices joined. You will see the impact of this when you study collision domains in a later chapter.

Modern Ethernet uses switches instead of hubs. With switches, each device normally has its own dedicated link to a port on the switch. The switch can forward different conversations at the same time over different ports. This makes collisions effectively disappear in standard full duplex environments and allows far better use of available bandwidth.

So, when people talk about Ethernet today, they almost always mean switched Ethernet, not the older shared medium style. Switching is what transformed Ethernet from a small office technology into the scalable foundation of enterprise and data center networks.

What a Switch Does Conceptually

At a high level, an Ethernet switch is an intelligent box that interconnects devices at the data link layer. It receives Ethernet frames on one port and decides which port or ports to send them out on so that the frame reaches the correct destination.

Inside a simple switched network, each device is plugged into a port on the switch. Devices use MAC addresses to identify the source and destination of Ethernet frames. The switch learns which MAC addresses are reachable on which ports and keeps that information in a table. Later in this section you will see in detail how this MAC learning works.

With this learned information, the switch can forward frames directly from source to destination. Frames are no longer broadcast to every device unless they really need to be, for example when the destination is unknown or the frame itself is a broadcast. This reduces unnecessary traffic and improves performance.

From the user point of view, the switch makes the local network feel like a single seamless environment where any device can talk to any other, as long as they share the same logical network. From a technical point of view, the switch is continuously receiving, processing, and re transmitting frames at very high speed.

How Switching Improves Network Performance

Switched Ethernet improves performance compared to shared media environments in several key ways. Each port typically has its own dedicated bandwidth, which means one busy device does not necessarily slow down others on different ports. The switch can also handle multiple conversations at once, for example one device talking to a server and another device talking to a printer, both at full speed.

The removal of collisions is another crucial improvement. When each device has its own full duplex link to the switch, it can send and receive at the same time without worrying about interfering with other devices. This is why you rarely hear about CSMA/CD in modern switched networks, even though it was central in early Ethernet.

Switches also segment the network for certain types of traffic. They restrict where most unicast frames travel and limit the scope of collisions. Broadcast traffic, however, still reaches all devices in the same broadcast domain. A later chapter in this section will explain the concepts of collision domains and broadcast domains in detail and show exactly how switches handle these.

As networks grow, switches also support advanced features that help maintain good performance, such as VLANs and Spanning Tree Protocol. VLANs allow traffic separation without separate physical hardware. Spanning Tree Protocol prevents loops in complex topologies that use multiple interconnected switches. Each of these topics will be covered in its own chapter.

Where Ethernet and Switching Are Used

Ethernet and switches appear almost everywhere inside local area networks. In a small office or home, a single switch may be built into the router that your Internet service provider supplies. All your wired devices plug into that one device, which acts as a small switch. Wireless access points often connect to this switch so that Wi Fi and wired clients can communicate over the same Ethernet network.

In larger offices, separate access switches serve groups of end devices on each floor or area. These access switches connect upwards to distribution switches, and those may connect further to core switches or routers. At every level, Ethernet links and switching move the frames between devices and between different parts of the network.

In data centers, high performance switches connect servers and storage systems with very high speed Ethernet links. In campus networks, switches connect phones, cameras, access points, and desktops. Even many industrial and IoT environments rely on ruggedized Ethernet switches as the basic wiring fabric.

The important point is that when you plug a network cable into a device inside a building, you are almost always entering an Ethernet switched environment, even if you do not see all the switches behind the scenes.

Relationship With Other Network Layers

Ethernet and switching operate at the lower levels of the networking stack. Ethernet frames wrap higher layer data, such as IP packets, which in turn carry transport protocols like TCP or UDP. The switch itself does not need to understand the application data. It focuses on MAC addresses and frame handling.

This separation of concerns is what allows Ethernet to act as a universal local transport for many different higher level protocols. Whether you are loading a web page, streaming video, sending email, or accessing a file share, the actual bits on the wire inside your local network are almost always organized into Ethernet frames and forwarded by switches.

Routing devices take over when traffic must move between different IP networks, often at the boundaries between VLANs or between your local network and the wider internet. This chapter does not cover routing in detail, but you should understand that switching focuses on moving frames within a broadcast domain, while routers handle movement between different networks.

Key Properties of Ethernet and Switching

Several core properties define modern Ethernet and switching and explain why they are so widely used.

First, Ethernet is scalable. Speeds have increased from 10 Mbps to 100 Gbps and beyond while keeping backward compatibility at the logical level. Network builders can upgrade parts of the network to higher speeds without changing how devices address each other or how frames are structured.

Second, Ethernet is flexible. It runs over multiple media types, such as copper and fiber, over short distances inside racks and long distances between buildings. The same switching concepts apply regardless of the medium.

Third, switching supports logical separation features. With VLANs, a single physical switch can host multiple logical networks. You can keep different departments or functions separated even if they plug into the same hardware. Trunking methods then carry multiple VLANs across single links between switches.

Fourth, modern switches are feature rich. They may support quality of service to prioritize important traffic, security features to control which devices can connect, monitoring capabilities, and redundancy functions like Spanning Tree and link aggregation to improve resilience and capacity.

All of these build on the basic idea of learning MAC addresses and forwarding Ethernet frames between ports.

Conceptual Flow of an Ethernet Frame Through a Switch

When a device wants to send data to another device on the same Ethernet network, it first wraps that data into an Ethernet frame. The frame includes a destination MAC address and a source MAC address, among other fields that will be explained later.

This frame travels over the physical cable to the switch port. The switch receives the frame, notes the source MAC address and the port it came in on, and records that association in its internal table. With this information it gradually builds a map of which devices are reachable on which ports.

Next, the switch checks the destination MAC address. If it already knows which port leads to that destination, it forwards the frame only out of that specific port. This is called unicast forwarding. If the switch does not know yet, it sends the frame out of all appropriate ports within that broadcast domain. This is called flooding. Once the destination device responds, the switch learns the location of that MAC address and can use unicast forwarding in future.

This simple loop of learning and forwarding allows switches to adapt automatically as devices connect, move, or disconnect. No manual configuration of MAC locations is needed in most basic environments, although advanced networks may apply additional policies on top.

Why Ethernet and Switching Matter for Beginners

For anyone starting in networking, understanding Ethernet and switching is essential because so many other topics build directly on them. When you later study MAC addresses, VLANs, and Spanning Tree in detail, you will be expanding ideas that start in this chapter.

Whenever you troubleshoot a local network problem, you are usually dealing with some part of this Ethernet and switching foundation. Is the switch port up? Is the device in the right VLAN? Is a broadcast storm affecting performance? These are practical questions that become understandable once you know what switches do with Ethernet frames.

Ethernet and switching also provide a concrete example of how layered networking works in real life. They show how physical links, data link addressing, and simple forwarding logic come together to produce the familiar experience of two computers on the same network seeing each other as if they were directly connected.

Important Concept Summary

To close this chapter, here are the central ideas that you should carry forward into the detailed subtopics that follow.

Understanding these basics will make it much easier to grasp MAC addresses, Ethernet frames, collision and broadcast domains, VLANs, trunking, and Spanning Tree Protocol in the upcoming chapters.

Views: 85