Skip to main content

What is Ethernet?

Ethernet is a networking technology and standard that defines how data is moved across cables in local area networks (LANs) and wide area networks (WANs). The Ethernet specification is governed by IEEE 802.3, established in 1983 and updated regularly. The Ethernet standard establishes how data is carried across the physical layer and data link layer. The standard covers data transmission rates, cabling and signaling types, and use of the Carrier Sense Multiple Access with Collision Detection (CSMA/CD) protocol. Ethernet is extensively used throughout organizations, buildings, and wider areas.

How does Ethernet work?

Ethernet was designed to solve a problem in data transmission: how can multiple devices communicate on a shared network without data packets colliding? Bob Metcalfe devised the solution, designing a system in which data packets are packaged into structures called frames. When sending frames, if the system detects two devices attempting to transmit at the same time, they both pause, wait a random interval, and try again.

This system means that data rarely overlaps in transmission, with the CSMA/CD (Carrier-Sense Multiple Access with Collision Detection) protocol detecting collisions and regulating transmission. Modern Ethernet now primarily uses full-duplex links and switches to make sure that data can be transmitted rapidly and without collisions, removing CSMA/CD. Ethernet operates on both Layer 1 (physical layer) and Layer 2 (Data link layer) of the Open Systems Interconnection (OSI) networking model.

The Open Systems Interconnection (OSI) Model

Physical layer

The physical layer of the OSI model manages the movement of electrical (or light for optical fiber cables) pulses across a cable. Different standards of Ethernet use different encoding techniques, but all standards similarly move data bits across the physical layer. The physical layer standardizes electrical signaling, modulation, encoding, and autonegotiation to establish a link between devices.

By running in full-duplex (where data can move in both directions concurrently), devices can both send and receive data without a chance of collision. Additional technologies within the physical layer, such as Forward Error Correction (FEC), help to maintain the quality of the signal transmissions, even at high speeds.

Data link layer

If the physical layer determines how bits move, then the data link layer manages where those bits go. Every network interface in the Ethernet system has a unique MAC (Media Access Controller) address (a 48-bit identifier), which allows switches to send the correct data packets to their destinations.

Instead of repeating signals across the whole network, the data layer allows for precise transmissions, isolating traffic and removing bottlenecks involved with sending signals out to the entire connected system.

The data link layer of the OSI model also implements flow control, helping to keep traffic steady with PAUSE frames.

The movement from hub-based topologies, where every single connected device received all transmissions, to the switching-based system enabled by MAC addresses was a major networking transformation. Switches that forward frames directly to MAC addresses allow the system to scale efficiently.

Ethernet frames

Frames, the small, structured Ethernet packets of data that move through local networks and segments, have a specific structure that allows them to travel. A frame includes several individual headers and trailers that help identify what the frame contains, where it came from, where it’s going, and what kind of data it carries. These include a start frame delimiter, destination MAC addresses, source MAC address, EtherType for data identification, and the actual payload of data.

What are the types of Ethernet?

Ethernet has evolved through many different standards, each of which brings new capabilities such as higher-speed data transmission. Here are some of the different types of Ethernet you might encounter.

Fast Ethernet

Originating in 1995 as IEEE 802.3u, Fast Ethernet can carry traffic at 100 Mbit/s, 10x faster than the previous version of Ethernet. Fast Ethernet uses twisted pair cables of copper wires, allowing companies that use previous versions of Ethernet to easily upgrade their existing systems and achieve this faster speed. This standard of Ethernet supports CSMA/CD in half-duplex mode, but also supports full-duplex transmission to avoid collisions.

Gigabit Ethernet standard

Three years after the arrival of Fast Ethernet, Gigabit Ethernet increased the theoretical transmission speeds of data over Ethernet to 1 gigabit per second. The main architectural change in the standard that allows for this increase in transmission speeds is line encoding. Both fiber optic and copper Ethernet variants have different line encoding strategies that allow for more bits per single signal.

Alongside line encoding changes in Layer 2, Gigabit Ethernet also uses all four wire pairs within copper cabling, again allowing organizations to upgrade their systems with little infrastructural change.

10 Gigabit and beyond

10 Gigabit Ethernet brought in a range of advanced changes that significantly improved speed transmission. First of all, 10GbE mainly uses fiber optic cables and 64b/66b encoding, reducing overhead and improving data transmission efficiencies. While the major gains of 10GbE are associated with fiber optics, more recent versions of Ethernet use FEC (Forward Error Correction) and digital signal processing (DSP) to retroactively allow copper versions to reach the same speeds.

Every generation of Ethernet since 10 Gigabit has continued to make physical-layer advancements, such as improving FEC (with 25 GbE) or using 4-level Pulse Amplitude Modulation (200G), to increase the spatial optimization of each Ethernet frame transmission.

What are the benefits of Ethernet?

There are several benefits of using Ethernet as a networking technology, helping to provide a secure, rapid, and effective way of transferring data.

Reliability and performance

When you need to rely on physical, local architecture, Ethernet is an excellent choice. Operating at Layer 1 and Layer 2, there are significantly fewer events that can interrupt an Ethernet connection, allowing for far fewer performance issues. Ethernet is stable and secure, making it an effective choice for applications that rely on low latency or need high-bandwidth data movement without errors.

Cost-effectiveness

Each major improvement to Ethernet represents significant gains in transmission speed, without changing the fundamental infrastructure of previous versions. This means organizations can upgrade their Ethernet without necessarily upgrading their physical hardware. Fast Ethernet to Gigabit and beyond still operate on familiar cabling structures and internal switching methods. Instead of having to completely replace network architecture when a new Ethernet standard comes out, organizations can improve their networks without hardware replacements, making Ethernet a highly cost-effective system.

Scalability

After the movement away from hub topologies, Ethernet became extremely scalable. Even a single switch can connect dozens of devices or endpoints, allowing communication at scale without having to significantly increase any of the underlying infrastructure. Modern versions of Ethernet also have scalability in mind, with virtual switching enabling organizations to implement Ethernet network segmentation without having to significantly change their systems.

Security

MACsec, IEEE 802.1AE, is a security standard that applies encryption directly to Ethernet transmission links, protecting any data in transit from hijacking or external interruption. By combining MACsec with Ethernet port security and additional Ethernet network access control features, Ethernet enables a highly secure form of data transmission.

What are the key characteristics of Ethernet?

Ethernet relies on a standardized system of physical infrastructure and connectors that provide the system with consistency and reliability. The basis of this system relies on cables (originally copper, but increasingly more commonly fiber optic), with copper cables like Cat5, Cat5e, and Cat6 offering a range of different speeds. Fiber cables are more costly to implement but also offer much higher speeds than copper alternatives.

Ethernet cables use connectors, which are components that link the physical cabling to network interfaces. Specific connectors, like RJ-45 for copper and LC for fiber, make sure that Ethernet switches and endpoints can intercommunicate without additional physical architecture.

Alongside this physical architecture, there are also Ethernet switches that forward frames to specific MAC addresses. The use of MAC helps to reduce the potential for collision in the system, improving its capacity to transmit information. Ethernet full-duplex design, where devices can send and receive concurrently, also supports scalability. Full-duplex replaced half-duplex transmission, which is where data could only flow in one direction at a time.

What is the difference between Ethernet and Wi-Fi?

Ethernet and Wi-Fi are both frameworks to transmit data, but the former uses Ethernet cables while the latter uses radio frequencies. Because Ethernet is a wired solution, it offers reliable speeds and low-latency transmission, but is also limited by the need for physical cabling architecture. Wi-Fi, by encoding data as radio waves and transmitting it digitally, is a more flexible system, but its performance depends on how far the data needs to travel and how much congestion there is on the network.

As Ethernet uses physical architecture, it is extremely safe, as an intruder needs to gain access to the cables or Ethernet ports to intercept traffic. WiFi uses several encryption standards to protect data in transmission, but this data must be encrypted to remain safe from listeners in range.

What are Ethernet network use cases?

Due to its reliability, scalability, and security, there are several use cases for Ethernet technology.

Data center networking

Data centers require predictable latency and high bandwidth to continually transfer large volumes of data. The evolution of Ethernet speeds makes Ethernet the standard for data center networking. Equally, as innovations within Ethernet standards offer backward compatibility with earlier versions, data centers have been able to easily update their networking to meet modern needs without major redesigns.

Enterprise networks

When physical devices reside in one place, such as workplace computers in an office, Ethernet cables offer a direct link to high-speed connectivity. For enterprise networks, either across major businesses or universities, for example, Ethernet is a good choice to supply predictable performance. Ethernet cables also offer a low-maintenance wired connection, helping businesses to provide their company with a network that can scale within the premises as needed.

Industrial applications

Many manufacturing systems still rely on Ethernet cable systems because they are considered to be more consistent than wireless connections. Network devices communicate through local networks, relying on the data link layer protocols to send signals at predictable intervals. This consistency, which can transmit data with minimal discrepancy, is useful for industrial applications that need real-time motion control or precise timing.

Cloud connectivity

Ethernet can be combined with cloud architecture, allowing wired local area networks to interconnect with a direct link to public cloud providers. A cloud data center contains thousands of network devices, ready to operate and intercommunicate at scale. Ethernet cable systems, unlike wireless connections, give these systems the consistent and reliable networking they need to handle high performance between on-premises and the cloud.

What are some best practices for Ethernet implementation?

Here are some best practices to consider in implementing Ethernet at scale.

Network design

Network design underpins effective Ethernet use. An example is deciding how to group network segments logically to make sure that multiple devices can communicate without bottlenecks. Where possible, choose compatible network interface card hardware and the most appropriate Ethernet protocol to provide long-term scalability and reliability.

Cable management

Cable management is the process of organizing, labeling, and routing the physical cables in your system and separating them to ensure you know which lines go to which part of your Ethernet service system. Alongside bringing some order to the mass of cables, effective management here also helps prevent signal interference and maintains system performance.

Redundancy

Any system that requires stability needs to consider how redundancy can help protect against any widespread failures or downtime. In an Ethernet system, businesses can plan for redundancy by offering multiple diverse pathways for signals, maintaining continuity even if one pathway is damaged or severed. Redundancy makes sure that multiple devices on the same system will continue to communicate, even if a network line fails.

Monitoring

As a general rule, continuous monitoring with network software helps you identify any interruptions, bottlenecks, or anomalies in your system that need further attention. Greater visibility over your system will also help point out where problems could be coming from, such as whether it’s internal to your network or a byproduct of upstream network services. Effective monitoring also ties into best practices for network security, because you can detect strange behavior more rapidly.

How can AWS support your Ethernet connectivity requirements?

AWS services are designed to connect with your on-premise Ethernet environments, for hybrid connectivity or to migrate to the cloud. Discover our range of networking services:

AWS Direct Connect allows you to create a dedicated network connection to AWS from your on-premises network, bypassing the public internet. AWS Direct Connect provides better performance, security, and connectivity to AWS.

Amazon Virtual Private Cloud (Amazon VPC) gives you full control over your virtual networking environment, including resource placement, connectivity, and security. You can set up your VPC, add resources, and define how your VPCs communicate with each other across accounts, Availability Zones, or AWS Regions.

AWS Transit Gateway connects your Amazon VPCs, AWS accounts, and on-premises networks to a single gateway. Transit Gateway acts as a highly scalable cloud router where each new connection is made only once.

Get started with networking on AWS by creating a free account today.

Browse all cloud computing concepts

Browse all cloud computing concepts content here:

Loading
Loading
Loading
Loading
Loading

Did you find what you were looking for today?

Let us know so we can improve the quality of the content on our pages