Ethernet Basics: Features, Devices, and Frame Structure

Ethernet is the most widely used technology in local wired networks. Almost all the LANs we work with today use Ethernet to connect systems and devices. In this tutorial, we’ll look at some important features of Ethernet, the devices used, and how the Ethernet frame is structured.

---

LAN

LAN stands for Local Area Network. It connects multiple devices like computers, printers, and switches within a small area such as a room, office, or lab.

• Devices in a LAN are usually connected using Ethernet cables or wireless.
• LANs don’t need an ISP. Devices can communicate with each other directly.
• Most LANs are built using switches and Ethernet-based communication.

---

Features of Ethernet

Ethernet is the technology used to send data between devices in a LAN. It uses MAC addresses and sends data in the form of Ethernet frames. Below are some of the main characteristics.

---

1. Connectionless Communication

Ethernet is a connectionless protocol. It doesn’t establish any connection before sending data. The sender just puts the frame on the wire. There is no handshake or session setup.

---

2. Unreliable Communication

Ethernet doesn’t guarantee delivery. If a frame is dropped or corrupted, it is not retransmitted. There’s no acknowledgment from the receiver. Reliability, if required, is handled by upper-layer protocols like TCP.

---

3. Unsecured Communication

Ethernet does not provide any security features. Any device on the same LAN can capture frames using tools like Wireshark. There is no encryption or authentication at the Ethernet level.

---

4. Simple, Fast, and Cost-Effective

Ethernet is commonly used in LANs because it’s fast, cheap, and easy to implement.

• Supports speeds like 100 Mbps, 1 Gbps, 10 Gbps and higher.
• Works over copper or fiber cables.
• Devices support auto-negotiation and are easy to set up.

---

Devices Used in Ethernet

Ethernet Hub

An Ethernet hub is one of the earliest devices used to connect multiple devices in a LAN. It’s a simple, physical-layer device that just takes in signals and sends them out to all other connected ports.

• If a PC sends a frame to the hub, the hub will forward that frame to all other connected devices.
• It doesn’t check or learn MAC addresses.
• It doesn’t know who the actual destination is.

Disadvantages of using hubs:

• Shared medium: All devices share the same bandwidth.
• Collision domain: If two devices transmit at the same time, collisions happen.
• Half duplex: Devices can either send or receive, not both at the same time.
• Very inefficient, especially when more devices are added.

That’s why hubs are no longer used in modern networks.

---

Ethernet Switch

An Ethernet switch solves all the major problems of hubs and is used in almost all Ethernet LANs today.

• It works at Layer 2, meaning it can read MAC addresses inside Ethernet frames.
• When a frame arrives at a switch, it checks the source MAC address and remembers which device is connected to which port.
• This information is stored in a MAC address table.

When the switch knows the destination MAC address:

• It sends the frame only to the correct port.

If the destination is unknown:

• It floods the frame to all ports except the one it came from (like a hub).

Advantages of switches over hubs:

• No collisions: Each port is a separate collision domain.
• Full duplex: Devices can send and receive at the same time.
• Efficient: Bandwidth is not wasted on devices that are not involved in the communication.
• Supports VLANs and other advanced features.

---

Ethernet Frame Format

The Ethernet frame contains the actual data being sent across the LAN. Different versions of Ethernet have slightly different frame structures.

Let’s go through the original DIX (DEC-Intel-Xerox) header and the updated IEEE 802.3 version.

---

Original Ethernet Header (DIX Format) aka “Ethernet II”.

The DIX header is the most commonly used version even today. It includes:

• Destination MAC Address (6 bytes): The device that should receive the frame.
• Source MAC Address (6 bytes): The device that sent the frame.
• Type (2 bytes): Indicates the protocol in the payload. Examples:
  • 0x0800 → IPv4
  • 0x0806 → ARP

This is followed by:

• Payload: Carries data from the upper layers (IP, ARP, etc.)
• FCS (4 bytes): Frame Check Sequence — used to detect errors.

Payload

Payload contains the actual data. Minimum size is 46 bytes. Maximum is 1500 bytes.
If the data is smaller than 46 bytes, padding is added to meet the minimum.

Trailer – FCS

At the end of the frame, there is a 4-byte FCS (Frame Check Sequence).

• It stores a CRC value calculated by the sender.
• Receiver recalculates CRC and compares it with FCS.
• If it doesn’t match, the frame is dropped. No retry.

This is how Ethernet detects errors, but it does not fix them.

---

IEEE 802.3 Header Format

IEEE later released an extended version that adds fields for more structured communication. The 802.3 frame includes:

• Destination MAC (6 bytes)
• Source MAC (6 bytes)
• Length (2 bytes): Size of the payload (up to 1500 bytes). This replaces the “Type” field in DIX.

Then comes the IEEE 802.2 LLC (Logical Link Control) header:

• DSAP (1 byte): Destination Service Access Point – identifies the protocol layer on the receiving device.
• SSAP (1 byte): Source Service Access Point – identifies the protocol layer on the sending device.
• Control (1 byte): Used for flow control or sequencing.

For networks that want to still use protocol identification (like IP), there’s an extension:

• OUI (3 bytes): Organizational Unique Identifier.
• Type (2 bytes): Same as the DIX “Type” field – specifies IPv4, ARP, etc.

This extended structure is used in environments where more layered control is needed.

---

Preamble and Start Frame Delimiter

Before any Ethernet frame begins, there’s a preamble to help the receiving device sync its clock with the sender’s.

• Preamble (7 bytes): A pattern of alternating 1s and 0s (10101010…)
• Start Frame Delimiter (1 byte): 10101011

This 8-byte sequence is sent before the actual frame and is not counted as part of the frame size. It’s mainly for synchronization.

---

Ethernet Speed and Naming Table

Different Ethernet versions are named based on speed, medium, and modulation. Here’s a table with common terms:

Speed	Standard Name	Transmission Medium	Notes
10 Mbps	10BASE-T	Twisted pair (copper)	Original Ethernet over UTP
100 Mbps	Fast Ethernet	100BASE-TX	Still used in older networks
1 Gbps	Gigabit Ethernet	1000BASE-T	Common in today’s LANs
10 Gbps	10-Gigabit Ethernet	10GBASE-T / 10GBASE-SR	Data centers, high-speed links
40 Gbps	40-Gigabit Ethernet	40GBASE-SR4	Short-range fiber links
100 Gbps	100-Gigabit Ethernet	100GBASE-LR4, etc.	Backbone links, providers

• “BASE” → Baseband transmission
• “T” → Twisted pair cable
• “SR/LR” → Short range / long range fiber

---