CS331 — Data Communications & Networking

Ethernet &
Local Area Networks

A complete study guide covering Ethernet standards, frame format, MAC addressing, sublayers, generations, and full-duplex operation.

01

What Is Ethernet?

📘 Definition

Ethernet is a LAN Data Link Layer protocol that is used in Bus and Star topologies and implements CSMA/CD (Carrier Sense Multiple Access with Collision Detection) as the medium access method.

Ethernet operates at the Data Link Layer (Layer 2) of the OSI model. It is the most widely used LAN technology in the world, connecting devices within a local area network using shared or switched media.

📅 Historical Timeline

🕰️ 1980 — Original Ethernet (DIX)

Developed by three companies:

  • Digital Equipment Corporation
  • Intel
  • Xerox

Together = DIX Ethernet (the original standard)

📋 1985 — IEEE Project 802

The Computer Society of IEEE started Project 802 to set standards enabling intercommunication among equipment from a variety of manufacturers.

Result: IEEE Ethernet (the current version)

🧠 Memory Trick — DIX

DIX = Digital + Intel + Xerox

Think: "Developed In eXcellence" — the three founders of original Ethernet in 1980.

🎯 Exam Focus

Know: Ethernet = LAN Data Link layer protocol, topologies = Bus & Star, access method = CSMA/CD, developed 1980 by DIX, standardized 1985 by IEEE Project 802.

📌 Section Summary
  • Ethernet is a Data Link Layer LAN protocol using CSMA/CD
  • Original Ethernet (1980) created by Digital, Intel, Xerox (DIX)
  • IEEE Project 802 (1985) standardized it → current IEEE Ethernet
  • Used in Bus and Star topologies
02

IEEE Ethernet — Two Sublayers

In IEEE 802.3 Ethernet, the Data Link Layer is split into two sublayers. This split allows Ethernet to be standardized in a way that different LAN protocols can share the same upper layer.

🔻 Bottom: MAC Sublayer

Standard: IEEE 802.3

  • Handles part of framing
  • Handles MAC addressing
  • Handles Medium Access Control
  • Specific implementation for each LAN protocol
  • Defines CSMA/CD for Ethernet LANs
  • Defines Token Passing for Token Ring

🔺 Top: LLC Sublayer

Standard: IEEE 802.2

  • Handles another part of framing
  • Provides error and flow control if needed
  • Makes the MAC sublayer transparent
  • Allows interconnectivity between different LAN data link layers
⚠️ Common Confusion — MAC vs LLC

MAC (802.3): Bottom sublayer. Technology-specific. Knows whether you are on Ethernet, Token Ring, etc. Handles the actual media access.

LLC (802.2): Top sublayer. Technology-independent. Provides a uniform interface to the network layer above. Makes different MACs look the same.

Think of it: MAC is the specialist (knows the hardware), LLC is the translator (hides the differences).

🎯 Exam Focus

MAC sublayer = IEEE 802.3 | LLC sublayer = IEEE 802.2

MAC is at the bottom; LLC is at the top of the Data Link layer.

LLC makes the MAC sublayer transparent, allowing interconnectivity between different LAN types.

📌 Section Summary
  • IEEE Data Link Layer is split into MAC (bottom) and LLC (top)
  • MAC = IEEE 802.3: framing, MAC addressing, medium access (CSMA/CD)
  • LLC = IEEE 802.2: error/flow control, makes MAC transparent
  • LLC allows different LAN technologies to communicate through a common interface
03

Ethernet Frame Format

There are two Ethernet frame formats: the original DIX Ethernet frame and the IEEE 802.3 frame. They are very similar — the key difference is that DIX uses a Type field while 802.3 uses a Length field.

📦 IEEE 802.3 Frame Structure (Visual)

Preamble
7 bytes
SFD
1 byte
Dest. Address
6 bytes
Source Address
6 bytes
Length/Type
2 bytes
Data & Padding
46–1500 bytes
CRC/FCS
4 bytes

🔍 Field-by-Field Explanation

Field Size Purpose & Details
Preamble 7 bytes (DIX: 8 bytes) Pattern 10101010 repeated. Used to synchronize sender and receiver clock rates before the actual frame arrives.
SFD
(Start Frame Delimiter)		 1 byte Pattern 10101011. Marks the start of the actual frame. Together with Preamble, forms the physical layer header.
Destination Address 6 bytes MAC address of the receiver. Can be Unicast, Multicast, or Broadcast.
Source Address 6 bytes MAC address of the sender. Always Unicast (a device can only send from its own address).
Type (DIX) 2 bytes Indicates the Network Layer protocol in the payload (e.g., IP=0800, ARP=0806, Novell IPX=8137, AppleTalk=809B). Value starts at 0600h (1536 decimal). Enables multiplexing.
Length (802.3) 2 bytes Number of bytes in the data field. Maximum = 1500 bytes (05DCh).
Data & Padding 46–1500 bytes Carries data from upper-layer protocols. Pad: zeros added if data is less than 46 bytes to ensure minimum frame size.
CRC / FCS 4 bytes Cyclic Redundancy Check (CRC-32). Checked at receiver — if error detected, frame is discarded. This is the trailer of the Data Link frame.
🎯 Exam Focus — Frame Length Limits

Minimum Frame Length

64 bytes

(512 bits)

Maximum Frame Length

1518 bytes

(12,144 bits)

Minimum data (payload) = 46 bytes | Maximum data = 1500 bytes

Why minimum 46 bytes? To satisfy CSMA/CD collision detection timing requirements. Pad zeros are added if data is smaller.

⚠️ Common Confusion — Type vs Length Field

DIX Ethernet: Uses a Type field (value ≥ 0600h = 1536 decimal) to identify the upper-layer protocol. Enables multiplexing of different network layer protocols.

IEEE 802.3: Uses a Length field (value ≤ 1500 = 05DCh) to specify how many bytes are in the data field.

How to tell which is which: If value > 1536 → Type field (DIX). If value ≤ 1500 → Length field (802.3).

🧠 Memory Trick — Frame Fields

"Pretty Students Do Seriously Long Data Calculations"

→ Preamble | SFD | Destination | Source | Length/Type | Data | CRC

📌 Section Summary
  • Preamble (7B) + SFD (1B) = Physical layer header for synchronization
  • Dest Address (6B) + Source Address (6B) = MAC addresses
  • Type (DIX, ≥0600h) or Length (802.3, ≤1500) = 2 bytes
  • Data: 46–1500 bytes; Pad added if <46 bytes
  • CRC-32 (4B) = error detection trailer; errors → frame discarded
  • Min frame: 64 bytes (512 bits) | Max frame: 1518 bytes (12,144 bits)
04

Ethernet (MAC) Address

📘 Definition — MAC Address

A MAC (Media Access Control) address is a 6-byte (48-bit) hardware address used to identify network devices at the Data Link layer. It is burned into the NIC ROM at the time of manufacture.

Structure of a MAC Address

Byte 1 — Byte 3
OUI — Vendor/Manufacturer ID
First 24 bits (3 bytes)
Byte 4 — Byte 6
Device ID — Unique per NIC
Last 24 bits (3 bytes)

Example address: 06-01-02-01-2C-4B

  • Size: 6 bytes = 48 bits
  • Format: Flat address — NOT hierarchical (unlike IP addresses)
  • Storage: Burned into NIC ROM (Read-Only Memory) — permanent
  • First 3 bytes: Identify the vendor/manufacturer (OUI — Organizationally Unique Identifier). Example: Cisco = 00-00-0C, Juniper = 00-05-85
  • Last 3 bytes: Created uniquely by the company for each device
  • Source address: ALWAYS Unicast — a device can only send from its own address
🎯 Exam Focus — Unicast vs Multicast Determination

Look at the least significant bit (LSB) of the first byte (= the second hexadecimal digit from the left):

  • If LSB = 0 (second hex digit is even) → Unicast
  • If LSB = 1 (second hex digit is odd) → Multicast
  • If all bits = 1 (all Fs: FF:FF:FF:FF:FF:FF) → Broadcast

📝 Worked Example

Given addresses — determine type:

Address 2nd Hex Digit Binary LSB Type
4A:30:10:21:10:1A A 1010 0 (even) ✅ Unicast
47:20:1B:2E:08:EE 7 0111 1 (odd) 📢 Multicast
FF:FF:FF:FF:FF:FF F 1111 All 1s 📡 Broadcast

Note: Look at the second hexadecimal character of the whole address (first hex char of byte 1). Even = Unicast; Odd = Multicast.

🧠 Memory Trick

"Even = One, Odd = Many"

Even second digit → Unicast (one recipient) | Odd second digit → Multicast (many recipients)

Broadcast is a special case of multicast where ALL bits are 1s (FF:FF:FF:FF:FF:FF)

📌 Section Summary
  • MAC address = 6 bytes (48 bits), flat (non-hierarchical), burned in NIC ROM
  • First 3 bytes = OUI (vendor ID), Last 3 bytes = unique device ID
  • Unicast: LSB of first byte = 0 (second hex digit even)
  • Multicast: LSB of first byte = 1 (second hex digit odd)
  • Broadcast: all bits = 1 → FF:FF:FF:FF:FF:FF (special case of multicast)
  • Source address is ALWAYS unicast
05

Types of Ethernet Destination Addresses

🎯 Unicast
LSB = 0

Sent to one specific station only. The second hex digit from the left is even.

Example: 4A:30:10:21:10:1A
(A = 1010 → LSB=0)

📢 Multicast
LSB = 1

Sent to a group of stations. The second hex digit from the left is odd. Used in conferencing applications.

Example: 47:20:1B:2E:08:EE
(7 = 0111 → LSB=1)

📡 Broadcast
All 1s

Sent to ALL stations on the network. Special case of multicast where all 48 bits = 1.

Always: FF:FF:FF:FF:FF:FF

💡 Important Rule

The source address is ALWAYS unicast. A device can only originate frames from its own unique MAC address — it cannot "pretend" to send from a group address.

🎯 Exam Focus — Key Notes

The least significant bit of the first byte determines Unicast (0) vs Multicast (1).

Broadcast = special case of multicast with all bits = 1 (FF:FF:FF:FF:FF:FF).

Multicast example use case: video conferencing where one station needs to send to a group simultaneously.

06

Ethernet Generations & Standards

Ethernet has evolved through four generations, each offering higher speeds while maintaining backward compatibility in addressing and frame format.

Standard Ethernet

10 Mbps

Original. Bus or Star topology.

Fast Ethernet

100 Mbps

10× faster. Star topology.

Gigabit Ethernet

1 Gbps

100× faster. Star topology.

10 Gigabit Ethernet

10 Gbps

Fiber only. Star topology.

Standard Ethernet — Naming Convention

🧠 How to Read Ethernet Names (e.g., 10Base-T)

10 = Speed in Mbps  |  Base = Baseband (digital)  |  T = Twisted pair cable, F = Fiber optic

So: 100Base-TX = 100 Mbps, Baseband, Twisted pair (category 5 UTP)

Detailed Standards Comparison

Standard Speed Topology Media Implementations
Standard Ethernet
10Base-T / 10Base-F		 10 Mbps Star (with switch) UTP or Fiber 10Base-T (2 pairs UTP), 10Base-F (2 fiber-optic cables)
Fast Ethernet
100Base-T / 100Base-F		 100 Mbps Star Twisted pair or Fiber 100Base-TX (2w Cat5 UTP), 100Base-FX (2w fiber), 100Base-T4 (4w Cat3 UTP)
Gigabit Ethernet
1000Base-T / 1000Base-F		 1000 Mbps (1 Gbps) Star Twisted pair or Fiber 1000Base-SX (short fiber), 1000Base-LX (long fiber), 1000Base-CX (STP copper), 1000Base-T (4w UTP)
10 Gigabit Ethernet
10GBase-F		 10,000 Mbps (10 Gbps) Star Fiber optic ONLY Fiber optic cables only — no copper support
🎯 Exam Focus — Key Differences Between Generations
  • Standard: 10 Mbps | Fast: 100 Mbps | Gigabit: 1 Gbps | 10G: 10 Gbps
  • All newer generations use Star topology
  • Only 10 Gigabit Ethernet uses fiber optic cable ONLY (no twisted pair option)
  • Fast Ethernet and Gigabit Ethernet support both twisted pair AND fiber
📌 Section Summary
  • Four generations: Standard (10M), Fast (100M), Gigabit (1G), 10 Gigabit (10G)
  • All use Star topology with a switch as central device
  • 10 Gigabit = fiber optic ONLY (no copper/UTP)
  • Naming: [speed]Base-[media] → T=twisted pair, F=fiber
  • Fast Ethernet implementations: 100Base-TX, 100Base-FX, 100Base-T4
  • Gigabit implementations: 1000Base-SX/LX (fiber), 1000Base-CX/T (copper)
07

Half Duplex vs Full Duplex Operation

📻 Half Duplex (Traditional)

  • Station can either transmit OR receive — not both simultaneously
  • Uses CSMA/CD to detect collisions
  • Traditional Ethernet operates this way
  • Central device: Hub
  • Shared collision domain

⚡ Full Duplex (Modern)

  • Station can transmit AND receive simultaneously
  • Throughput is doubled
  • No collisions → CSMA/CD NOT needed
  • Central device: Switch (not hub!)
  • Each station has its own collision domain

Full Duplex Throughput Examples

100 Mbps Ethernet in Full-Duplex

→ 200 Mbps

theoretical transfer rate

1 Gbps Ethernet in Full-Duplex

→ 2 Gbps

theoretical transfer rate

Requirements to Enable Full-Duplex Mode

To operate in full-duplex mode, 4 changes must be made:

# Requirement Why?
1 Stations must have full-duplex NIC cards To physically support simultaneous send/receive
2 Must use two pairs of wire One pair: host → switch (inbound). Other pair: switch → host (outbound)
3 Must use a Switch as central device (NOT a hub) Switches create dedicated connections; hubs broadcast to all
4 Devices must connect point-to-point (dedicated) to the switch Dedicated connection = no shared medium = no collisions
⚠️ Common Confusion — Hub vs Switch in Duplex

Hub: Broadcasts all traffic to all ports. Shared medium → collisions can happen → CSMA/CD needed → half duplex only.

Switch: Creates a dedicated path between sender and receiver. Each port is its own collision domain → no collisions → CSMA/CD not needed → can operate in full duplex.

Full duplex requires a switch — never a hub.

🎯 Exam Focus

In full-duplex: (1) Throughput is doubled. (2) CSMA/CD is NOT needed because there are no collisions. (3) Each station constitutes a separate collision domain. (4) Must use a switch, not a hub.

📌 Section Summary
  • Half duplex: transmit OR receive (not both) — uses CSMA/CD, hub-based
  • Full duplex: transmit AND receive simultaneously — doubles throughput
  • Full duplex requirements: full-duplex NIC, 2 wire pairs, switch (not hub), point-to-point connection
  • In full duplex: no collisions → CSMA/CD algorithm not needed
  • Each station in switched Ethernet = separate collision domain
08

⚡ Quick Revision Cheat Sheet

Ethernet Basics

LayerData Link (2)
Access MethodCSMA/CD
TopologiesBus & Star
Created1980 — DIX
Standardized1985 — IEEE 802

Frame Fields & Sizes

Preamble7 bytes
SFD1 byte (10101011)
Dest / Src Address6 bytes each
Type / Length2 bytes
Data46–1500 bytes
CRC4 bytes (CRC-32)

Frame Limits

Min frame64 bytes / 512 bits
Max frame1518 bytes
Min data46 bytes
Max data1500 bytes
Pad added whendata < 46 bytes

MAC Address

Size6 bytes / 48 bits
TypeFlat (not hierarchical)
StorageNIC ROM (permanent)
First 3 bytesVendor (OUI)
Last 3 bytesUnique device ID

Address Types

UnicastLSB=0, even digit
MulticastLSB=1, odd digit
BroadcastFF:FF:FF:FF:FF:FF
Source addressAlways Unicast

Ethernet Generations

Standard10 Mbps
Fast100 Mbps
Gigabit1 Gbps
10 Gigabit10 Gbps (fiber only)

Sublayers

LLC (Top)IEEE 802.2
LLC handlesError/flow control
MAC (Bottom)IEEE 802.3
MAC handlesFraming, CSMA/CD

Full Duplex

ThroughputDoubled
CSMA/CDNot needed
Central deviceSwitch (not hub)
Wire pairsTwo pairs
ConnectionPoint-to-point
🧠 All Key Numbers at a Glance
  • MAC address = 6 bytes = 48 bits
  • Preamble = 7 bytes (pattern 10101010)
  • SFD = 1 byte (pattern 10101011)
  • Type field starts at 0600h = 1536 decimal
  • Max Length field = 1500 bytes = 05DCh
  • Min payload = 46 bytes (pad if less)
  • Max payload = 1500 bytes
  • Min frame = 64 bytes, Max = 1518 bytes
09

🧪 Practice Questions

Q1Ethernet was originally developed in 1980 by which three companies?
✅ Correct: B. The original Ethernet was developed in 1980 by Digital, Intel, and Xerox — together abbreviated as DIX. IEEE standardized it in 1985 via Project 802.
Q2What is the MAC sublayer standard number, and what access method does it define for Ethernet LANs?
✅ Correct: C. The MAC sublayer uses standard IEEE 802.3 and defines CSMA/CD as the access method for Ethernet LANs. (Token Passing is for Token Ring, not Ethernet.) The LLC sublayer uses IEEE 802.2.
Q3What is the minimum and maximum size of the data field in an Ethernet frame?
✅ Correct: B. The data field must be between 46 and 1500 bytes. If the data is smaller than 46 bytes, padding (zeros) is added to meet the minimum. The total frame (including all headers) is 64–1518 bytes.
Q4Determine the type of this destination address: 47:20:1B:2E:08:EE
✅ Correct: B. Look at the second hexadecimal digit of the address: 7. Convert 7 to binary: 0111. The LSB (least significant bit) = 1, which means it is Multicast. Odd second digit = Multicast.
Q5A MAC address has its first 3 bytes as 00-00-0C. What does this represent?
✅ Correct: B. The first 3 bytes (24 bits) of a MAC address identify the vendor/manufacturer (called the OUI — Organizationally Unique Identifier). 00-00-0C belongs to Cisco. The last 3 bytes are the unique device identifier assigned by the manufacturer.
Q6What is the difference between the Type field (DIX) and the Length field (IEEE 802.3)?
✅ Correct: C. In DIX Ethernet, the Type field (value ≥ 0600h = 1536) identifies the upper-layer protocol (e.g., IP=0800, ARP=0806). In IEEE 802.3, the Length field (value ≤ 05DCh = 1500) specifies how many bytes are in the data field. The value distinguishes which format is used.
Q7Which of the following is NOT a requirement for full-duplex Ethernet operation?
✅ Correct: C. Full duplex requires a switch (NOT a hub) as the central device. A hub broadcasts traffic to all ports creating collisions, making full duplex impossible. A switch creates dedicated connections between ports. All other options (full-duplex NIC, two wire pairs, point-to-point connection) are indeed requirements.
Q8Which Ethernet generation uses fiber optic cable ONLY (no copper/UTP support)?
✅ Correct: D. 10 Gigabit Ethernet uses fiber optic cable only. Standard Ethernet, Fast Ethernet, and Gigabit Ethernet all support both twisted pair (copper) and fiber optic. The very high speed of 10 Gbps requires the bandwidth of fiber and cannot reliably run over copper at that distance.