3.3 Ethernet Frame Format (IEEE 802.3) & Minimum Frame Size

Back to Computer Networks Notes

📚Module 03: Data Link Layer: MAC Sublayer & EthernetTopic 3.3⏱️12 min read

🎯High-Yield For:Semester Exams (All Universities) • Placement Technical Rounds • Engineering Foundations

💡 Core Intuition: The Postal Envelope Analogy

🍳 The Everyday Analogy: The Standard Certified Mail Envelope

Imagine sending a letter via postal service:

1. The Preamble & Wake-up Bell: Before handing the letter to the sorter, you ring a loud brass bell for 7 seconds (10101010...) so the clerk stops what they are doing and tunes their ears to your rhythm. When you abruptly change tone (10101011), the clerk knows the exact letter text starts now.
2. Addresses & Metadata: On the front of the envelope, you write the recipient's permanent residential address (6-byte Destination MAC) and your own return address (6-byte Source MAC).
3. The Minimum Size Rule (Padding): Postal sorting machines have conveyor belts with rollers spaced 10 cm apart. If you try to mail a tiny postage stamp of paper alone, it falls between the rollers and jams the machine! To mail even a 1-word note, the post office requires you to place it inside a standard $64\text{-byte}$ envelope, adding cardboard spacers (Padding) if your note is too small.
4. The Tamper Seal: At the back flap, you stamp a wax cryptographic seal (4-byte CRC-32). If the seal is cracked upon arrival, the entire envelope is discarded into the furnace.

💻 Bridging to Computer Science

Created by Robert Metcalfe and David Boggs at Xerox PARC in 1973 and standardized as IEEE 802.3, Ethernet is the dominant Local Area Network (LAN) architecture in human history.

Every packet transmitted by an operating system down through the TCP/IP stack is wrapped in an Ethernet Frame. Understanding each byte of this frame—from physical clock synchronization down to hardware error checksums—is essential for understanding real-world networking.

---

📑Table of Contents

• 🔀 Ethernet Frame Architecture (IEEE 802.3 vs Ethernet II)
• 🔬 Field-by-Field Frame Anatomy
  • 1. Preamble (7 Bytes / 56 Bits)
  • 2. Start Frame Delimiter (SFD) (1 Byte / 8 Bits)
  • 3. Destination MAC Address (6 Bytes / 48 Bits)
  • 4. Source MAC Address (6 Bytes / 48 Bits)
  • 5. Length / EtherType (2 Bytes / 16 Bits)
  • 6. Data Payload (46 to 1500 Bytes)
  • 7. Cyclic Redundancy Check (CRC-32 / FCS) (4 Bytes / 32 Bits)
• 🏷️ 48-Bit MAC Addresses: Unicast, Multicast & Broadcast
  • Unicast vs Multicast vs Broadcast Detection
• 📏 The 64-Byte Minimum Frame Size Derivation
  • Derivation from CSMA/CD Physics
  • Why Minimum Payload is Exactly 46 Bytes:
• 📦 The 1500-Byte Maximum Transmission Unit (MTU)
• ⚡ Physical Layer Signaling: Manchester Encoding
  • Why Manchester Encoding?
• ⏱️ Wire Footprint: Preamble, SFD & Inter-Packet Gap (IPG)
• 🏭 Real-World Production Context: Jumbo Frames & 802.1Q VLAN Tagging in Modern Data Centers
  • 1. Jumbo Frames (9000-Byte MTU) in SAN & Storage Clusters
  • 2. IEEE 802.1Q VLAN Tagging (Frame Expansion to 1522 Bytes)
• 📐 Numerical Applications & Worked Examples
  • Example 1: Wire-Speed Forwarding Packet Processing Time
  • Example 2: Maximum Cable Length for 500 Mbps Ethernet
  • Example 3: Minimum Frame Size with Jamming Signal Delay
  • Example 4: Unicast vs Multicast Address Classification
• 🎯 Exam & Interview Pitfall Check

---

🔀 Ethernet Frame Architecture (IEEE 802.3 vs Ethernet II)

Two primary Ethernet frame standards exist: the original Ethernet II (DIX) standard (developed by Digital Equipment, Intel, and Xerox) and the formal IEEE 802.3 international standard.

Comparing the two coexisting Ethernet framing standards

Industry Standard

Ethernet II (DIX Standard)

🌐

Dominant Architecture / DomainModern Internet & TCP/IP Traffic

• •Uses a 2-byte 'Type' field (EtherType) indicating the enclosed Network Layer protocol.
• •EtherType values are >= 1536 (0x0600), e.g. 0x0800 for IPv4, 0x0806 for ARP, 0x86DD for IPv6.
• •Eliminates sublayer LLC/SNAP overhead, allowing native IP packet encapsulation.
• •Represents over 99% of all traffic on the modern Internet.

"Direct protocol multiplexing via 16-bit EtherType values."

Formal IEEE Standard

IEEE 802.3 Standard

📜

Dominant Architecture / DomainLegacy & Enterprise Protocol Stacks

• •Replaces the 'Type' field with a 2-byte 'Length' field indicating payload byte count.
• •Length values are strictly <= 1500 (0x05DC).
• •Demultiplexing is delegated to the upper LLC sublayer (IEEE 802.2 SAP or SNAP headers).
• •Coexists seamlessly on the same wire because Length (<= 1500) and Type (>= 1536) never overlap!

"Formal length demarcation with higher-layer LLC multiplexing."

---

🔬 Field-by-Field Frame Anatomy

IEEE 802.3 / Ethernet II StandardTotal Frame: 64 to 1518 Bytes

Frame Format Anatomy

Hover or click any field to inspect byte allocations, physical bit patterns, and protocol rules

PreamblePHY

7 Bytes

10101010...

SFDPHY

1 Byte

10101011

Dest MACMAC

6 Bytes

48 Bits

Source MACMAC

6 Bytes

48 Bits

Length / TypeMAC

2 Bytes

16 Bits

Data PayloadLLC/Net

46 - 1500 Bytes

MTU = 1500B

CRC-32 (FCS)MAC

4 Bytes

32-Bit FCS

7 BytesPHY Sublayer10101010...Preamble

Generates a 10 MHz square wave allowing the receiver's Phase-Locked Loop (PLL) clock circuit to lock frequency and synchronize with the incoming bit stream.

Hardware Rule:Generated and stripped off by the physical layer transceiver; not counted in official MAC frame byte count.

1. Preamble (7 Bytes / 56 Bits)

• Bit Pattern: Alternating sequence of ones and zeros: 10101010 10101010 ... 10101010 ($7 \times 8 = 56\text{ bits}$).
• Purpose: Generates a periodic alternating $10\text{ MHz}$ square-wave clock signal. When bits hit an idle copper wire, the receiver's physical phase-locked loop (PLL) circuitry locks its clock frequency to the transmitter's bit clock.
• Note: The Preamble is generated and stripped off by the physical layer transceiver; it is not counted in the official frame size.

2. Start Frame Delimiter (SFD) (1 Byte / 8 Bits)

• Bit Pattern: 10101011.
• Purpose: The pattern continues the alternating rhythm until the final two consecutive ones (11). These two consecutive ones break the rhythm and alert the receiver's hardware: "The synchronization preamble has ended; the very next bit is the first bit of the Destination MAC address."

3. Destination MAC Address (6 Bytes / 48 Bits)

• Contains the hardware physical address of the receiving interface.
• Present on the target Network Interface Card (NIC).

4. Source MAC Address (6 Bytes / 48 Bits)

• Contains the hardware physical address of the transmitting interface.
• Crucial Rule: The Source MAC address is always a Unicast address; a broadcast or multicast address can never generate a frame.

5. Length / EtherType (2 Bytes / 16 Bits)

• If Value $\le 1500$ ($0\text{x}05\text{DC}$): Interpreted as IEEE 802.3 Length, specifying the exact number of data payload bytes.
• If Value $\ge 1536$ ($0\text{x}0600$): Interpreted as Ethernet II EtherType:
  • 0x0800: IPv4 Datagram
  • 0x0806: Address Resolution Protocol (ARP)
  • 0x86DD: IPv6 Datagram
  • 0x8100: IEEE 802.1Q VLAN Tag

6. Data Payload (46 to 1500 Bytes)

• Encapsulates the upper-layer packet (e.g. an IPv4 packet).
• Minimum Payload Constraint: $46\text{ bytes}$. If the upper layer passes a payload smaller than 46 bytes (such as a 28-byte ARP packet), the MAC layer appends Padding bytes (zeros) to reach exactly 46 bytes.
• Maximum Payload Constraint: $1500\text{ bytes}$ (Standard Ethernet Maximum Transmission Unit - MTU).

7. Cyclic Redundancy Check (CRC-32 / FCS) (4 Bytes / 32 Bits)

• Frame Check Sequence (FCS) calculated using the standard 32-bit CRC polynomial: $G(x) = x^{32} + x^{26} + x^{23} + x^{22} + x^{16} + x^{12} + x^{11} + x^{10} + x^8 + x^7 + x^5 + x^4 + x^2 + x + 1$
• Calculated over all fields from Destination MAC through Data/Padding.
• If the receiver calculates a remainder different from zero, it discards the frame immediately. Ethernet provides error detection, not error correction.

---

🏷️ 48-Bit MAC Addresses: Unicast, Multicast & Broadcast

A Media Access Control (MAC) address is a globally unique 48-bit identifier burned into the read-only memory (ROM) of every network interface card:

Field Segment	Bit Count / Byte Count	Assigned Authority & Role
Organizationally Unique Identifier (OUI)	First 24 Bits (3 Bytes)	Globally assigned by IEEE Registration Authority to hardware vendors (e.g. Cisco, Intel, Apple).
Network Interface Controller (NIC) Specific	Last 24 Bits (3 Bytes)	Serial number assigned uniquely by the hardware manufacturer to the specific physical interface.

1. Organizationally Unique Identifier (OUI - First 24 Bits): Assigned by the IEEE to hardware manufacturers (e.g. Cisco, Apple, Intel).
2. Network Interface Controller Specific (NIC - Last 24 Bits): Serial number assigned by the vendor.

Unicast vs Multicast vs Broadcast Detection

How does hardware determine whether a frame is destined for a single node, a group, or everyone on the LAN? By inspecting the Least Significant Bit (LSB) of the very first byte transmitted on the wire:

$\text{First Byte Binary: } b_7 b_6 b_5 b_4 b_3 b_2 b_1 \mathbf{b_0}$

Address Type	Least Significant Bit ($\mathbf{b_0}$)	Description & Examples
Unicast	$\mathbf{0}$	Frame destined for a single unique host. E.g. 00:1A:2B:... (00 in hex is 00000000_2, LSB is 0).
Multicast	$\mathbf{1}$	Frame destined for a registered group of hosts. E.g. 01:00:5E:... (01 in hex is 00000001_2, LSB is 1).
Broadcast	All 48 bits = $\mathbf{1}$	Frame processed by every active station on the LAN: FF:FF:FF:FF:FF:FF.

Universally vs Locally Administered Bit ($\mathbf{b_1}$): The second least significant bit ($b_1$) of the first byte indicates administrative scope: $0 = \text{Globally Unique (IEEE OUI)}$, $1 = \text{Locally Administered (configured by network administrator or VM hypervisor)}$.

---

📏 The 64-Byte Minimum Frame Size Derivation

Why is the minimum length of an Ethernet frame strictly 64 bytes (512 bits)?

Frame Field	Length	CSMA/CD Operational Role
Destination MAC Address	$6\text{ Bytes}$	Identifies physical target interface
Source MAC Address	$6\text{ Bytes}$	Identifies transmitting physical interface
Length / EtherType	$2\text{ Bytes}$	Demarcates protocol payload format
Minimum Data Payload	$46\text{ Bytes}$	Minimum required data (padding added if smaller)
CRC-32 Checksum (FCS)	$4\text{ Bytes}$	Frame error check sequence
Total Minimum Frame Size	$\mathbf{64\text{ Bytes}}$	512 Bits (satisfies $T_t \ge 2 T_p$ at 10 Mbps)

Derivation from CSMA/CD Physics

As derived in Topic 3.2, for a transmitting station to reliably detect a collision before it finishes transmitting, its transmission time ($T_t$) must be at least twice the worst-case propagation delay of the cable ($2 T_p$):

$T_t \ge 2 \times T_p \implies \frac{L_{\min}}{B} \ge 2 T_p \implies \Large L_{\min} = 2 \times T_p \times B$

In the original IEEE 802.3 specification:

• Network Bandwidth: $B = 10\text{ Mbps} = 10^7\text{ bps}$.
• Maximum collision domain diameter (5 segments of $500\text{ m}$ thick coaxial cable connected by 4 repeaters plus transceiver cables): $\approx 2500\text{ meters}$.
• Round-trip propagation delay including repeater signal regeneration delays: $2 T_p \approx 51.2\text{ microseconds}$

Substituting these physical values:

$L_{\min} = (51.2 \times 10^{-6}\text{ s}) \times (10 \times 10^6\text{ bps}) = \mathbf{512\text{ bits}}$

$\text{Minimum Frame Size in Bytes} = \frac{512\text{ bits}}{8\text{ bits/byte}} = \mathbf{64\text{ bytes}}$

Why Minimum Payload is Exactly 46 Bytes:

$\text{Fixed Frame Header and Trailer} = \underbrace{6}_{\text{Dest MAC}} + \underbrace{6}_{\text{Src MAC}} + \underbrace{2}_{\text{Type}} + \underbrace{4}_{\text{CRC}} = 18\text{ bytes}$

$\text{Minimum Payload} = \text{Total Minimum} - \text{Header/Trailer} = 64\text{ bytes} - 18\text{ bytes} = \mathbf{46\text{ bytes}}$

If an upper-layer application generates a message smaller than 46 bytes (such as a 20-byte TCP acknowledgment or an ARP request), the network interface card automatically injects dummy padding bytes until the payload reaches 46 bytes.

---

📦 The 1500-Byte Maximum Transmission Unit (MTU)

The Maximum Transmission Unit (MTU) of Standard Ethernet is 1500 bytes, setting the maximum total frame size to:

$\text{Maximum Frame Size} = 18\text{ bytes (Headers + CRC)} + 1500\text{ bytes (Payload)} = \mathbf{1518\text{ bytes}}$ (Or $1522\text{ bytes}$ if an IEEE 802.1Q 4-byte VLAN tag is present).

Why didn't Ethernet's designers allow unbounded frame sizes?

1. Buffer Memory Cost (1970s Economics): In 1973, high-speed RAM buffers were extraordinarily expensive. If frames could be arbitrarily large (e.g. 100 KB), every simple network card would require huge on-board memory chips to buffer incoming frames, drastically increasing equipment cost.
2. Channel Monopolization: In a shared broadcast bus, if one host begins transmitting a 50 MB file as a single uninterrupted frame, all other nodes are blocked from the wire for seconds, causing massive transmission jitter for real-time traffic.
3. Bit Error Vulnerability: Over noisy physical copper, the probability of a frame surviving without a single bit error decreases exponentially with frame length ($P_{\text{clean}} = (1 - \text{BER})^L$). If an error occurs in a huge frame, the entire massive frame must be discarded and retransmitted, severely wasting bandwidth.

---

⚡ Physical Layer Signaling: Manchester Encoding

Standard $10\text{ Mbps}$ Ethernet (10BASE5, 10BASE-T) does not use simple Non-Return-to-Zero (NRZ) voltage levels. It uses Manchester Encoding:

Binary Bit	Physical Center Transition	Signal Representation
Bit 0	High-to-Low ($\downarrow$)	Voltage drops from $+V$ to $-V$ at the exact center of the bit duration.
Bit 1	Low-to-High ($\uparrow$)	Voltage rises from $-V$ to $+V$ at the exact center of the bit duration.

Why Manchester Encoding?

1. Self-Clocking Synchronization: Because every single bit has a voltage transition dead in the center, the receiver recovers the sender's clock continuously on every bit interval. Long strings of consecutive zeros or ones never cause clock drift.
2. Zero DC Baseline Wander: The signal spends equal time at positive and negative voltages, maintaining a net zero DC offset across isolation transformers.
3. Baud Rate Trade-off: Because there are two signal states (transitions) per data bit: $\text{Baud Rate (Modulation Rate)} = 2 \times \text{Bit Rate}$ A $10\text{ Mbps}$ Ethernet link requires a physical signaling rate of 20 Megabaud (20 MHz), yielding a spectral efficiency of $50\%$.

---

⏱️ Wire Footprint: Preamble, SFD & Inter-Packet Gap (IPG)

When an Ethernet frame is transmitted on the physical wire, the total channel occupancy includes the physical preamble, SFD, and a mandated quiet time between packets called the Inter-Packet Gap (IPG):

Physical Wire OccupancyTotal Frame: 84 Bytes (672 Bits)

Frame Format Anatomy

Total channel time allocation including physical synchronization and inter-packet silence

Preamble

7 Bytes

SFD

1 Byte

10101011

MAC Frame

64 Bytes

Min Frame

Inter-Packet Gap

12 Bytes

96 Bit Times

7 BytesPreamble

56-bit alternating 10 MHz clock sync square wave.

• Inter-Packet Gap (IPG): A mandatory silent pause of 96 bit times between consecutive frames ($9.6\text{ }\mu\text{s}$ at $10\text{ Mbps}$, $96\text{ ns}$ at $100\text{ Mbps}$, $9.6\text{ ns}$ at $1\text{ Gbps}$). It gives receiver electronics sufficient time to reset their internal buffers and line cards before the next frame arrives.

$\text{Minimum Wire Footprint} = \underbrace{7}_{\text{Preamble}} + \underbrace{1}_{\text{SFD}} + \underbrace{64}_{\text{Min Frame}} + \underbrace{12}_{\text{IPG}} = \mathbf{84\text{ Bytes (672 bits)}}$

---

🏭 Real-World Production Context: Jumbo Frames & 802.1Q VLAN Tagging in Modern Data Centers

In high-throughput enterprise infrastructure and hyper-scale cloud environments (AWS, Google Cloud, Azure), traditional standard Ethernet limits create severe operational challenges:

1. Jumbo Frames (9000-Byte MTU) in SAN & Storage Clusters

At modern line rates of $10\text{ Gbps}$, $40\text{ Gbps}$, and $100\text{ Gbps}$, transmitting data in traditional $1500\text{-byte}$ MTU chunks creates an enormous packet rate:

• A saturated $10\text{ Gbps}$ link pushes approximately 812,000 frames per second with a standard $1500\text{-byte}$ MTU.
• Every single frame triggers a hardware interrupt, network stack context switch, and CPU cycle overhead on host servers.
• The Jumbo Frame Solution: Storage Area Networks (iSCSI, NFS) and distributed database clusters enable Jumbo Frames (MTU 9000 bytes). At 9000 bytes, the frame rate drops to 138,000 frames per second—a nearly $6\times$ reduction in CPU interrupt frequency—dramatically reducing server CPU utilization and latency while maximizing disk-to-network pipeline throughput.

2. IEEE 802.1Q VLAN Tagging (Frame Expansion to 1522 Bytes)

In production networks, thousands of isolated virtual networks share physical switch infrastructure using IEEE 802.1Q Virtual LANs (VLANs):

• A 4-byte VLAN tag (containing a 12-bit VLAN ID allowing 4094 isolated subnets) is inserted directly between the Source MAC address and the EtherType field.
• This expands the maximum standard frame from $1518\text{ bytes}$ to $1522\text{ bytes}$. Modern switches are specially engineered with "Baby Giant" frame support to forward these tagged frames without dropping them as oversized runts.

---

📐 Numerical Applications & Worked Examples

Example 1: Wire-Speed Forwarding Packet Processing Time

Problem Statement: A core network router has two full-duplex Ethernet interfaces, each operating at $100\text{ Mbps}$. Ethernet frames on the physical wire are at least $84\text{ bytes}$ long (including the 7-byte Preamble, 1-byte SFD, 64-byte Minimum Frame, and 12-byte Inter-Packet Gap). Calculate the maximum packet processing time available at the router CPU for "wire-speed forwarding" (zero queue buildup) to be sustained under worst-case line rate.

Step-by-Step Solution:

1. Identify the Worst-Case Scenario: The router experiences maximum packet arrival rate when all arriving frames are of minimum size ($84\text{ bytes}$ physical footprint).
2. Calculate Bit Time at 100 Mbps: $\text{Bit Time} = \frac{1}{100 \times 10^6\text{ bps}} = 10\text{ nanoseconds} = 0.01\text{ }\mu\text{s}$
3. Calculate Time to Receive One Minimum-Sized Frame: $\text{Total Bits on Wire} = 84\text{ bytes} \times 8\text{ bits/byte} = 672\text{ bits}$ $\text{Frame Arrival Duration} = 672\text{ bits} \times 10\text{ ns/bit} = 6720\text{ nanoseconds} = \mathbf{6.72\text{ microseconds}}$
4. Conclusion: A new frame can arrive every $6.72\text{ }\mu\text{s}$. To forward packets at full wire speed without overflowing its input queues, the router must inspect the header, make the forwarding table lookup, and dispatch the packet in less than or equal to $6.72\text{ microseconds}$.

---

Example 2: Maximum Cable Length for 500 Mbps Ethernet

Problem Statement: An enterprise designs an Ethernet LAN operating at $500\text{ Mbps}$ with a standard frame size of $10,000\text{ bits}$. The signal propagation velocity in the cable is $200,000\text{ km/s}$ ($2 \times 10^8\text{ m/s}$). Determine the maximum allowable cable length (in kilometers) to preserve CSMA/CD collision detection.

Step-by-Step Solution:

1. Apply the CSMA/CD Condition: $T_t \ge 2 \times T_p \implies \frac{L}{B} \ge 2 \times \frac{d}{v}$
2. Rearrange for Maximum Distance ($d$): $d \le \frac{L \times v}{2 \times B}$
3. Substitute Values:
   • $L = 10,000\text{ bits} = 10^4\text{ bits}$
   • $v = 200,000\text{ km/s} = 2 \times 10^5\text{ km/s}$
   • $B = 500\text{ Mbps} = 500 \times 10^6\text{ bps} = 5 \times 10^8\text{ bps}$
4. Compute Distance: $d = \frac{10^4 \times (2 \times 10^5\text{ km/s})}{2 \times (5 \times 10^8\text{ bps})} = \frac{2 \times 10^9}{10 \times 10^8} = \frac{2 \times 10^9}{10^9} = \mathbf{2\text{ km}}$

---

Example 3: Minimum Frame Size with Jamming Signal Delay

Problem Statement: Suppose the round-trip propagation delay for a $10\text{ Mbps}$ Ethernet bus is $46.4\text{ microseconds}$. In addition, upon detecting a collision, a transmitting station emits a $48\text{-bit}$ jamming signal. Calculate the exact minimum frame size (in bits) required to guarantee collision detection before transmission finishes.

Step-by-Step Solution:

1. Understand the Jamming Signal Addition: The station must continue transmitting its frame until the returning collision arrives ($2 T_p$) PLUS the duration required to emit the jamming signal ($T_{\text{jam}}$): $T_t \ge 2 T_p + T_{\text{jam}}$ $\frac{L_{\min}}{B} = 2 T_p + \frac{L_{\text{jam}}}{B} \implies L_{\min} = 2 T_p \times B + L_{\text{jam}}$
2. Calculate Bits for Round-Trip Delay: $2 T_p \times B = (46.4 \times 10^{-6}\text{ s}) \times (10 \times 10^6\text{ bps}) = 46.4 \times 10 = 464\text{ bits}$
3. Add Jamming Signal Bits: $L_{\min} = 464\text{ bits} + 48\text{ bits} = \mathbf{512\text{ bits}}$ $\text{In Bytes} = \frac{512}{8} = \mathbf{64\text{ bytes}}$ (Notice how the IEEE 802.3 committee's exact 512-bit standard accounts for both 464 bits of physical round-trip propagation and 48 bits of jam signaling!)

---

Example 4: Unicast vs Multicast Address Classification

Problem Statement: Classify each of the following four MAC addresses as Unicast, Multicast, or Broadcast:

1. E2:45:67:89:AB:CD
2. 4B:23:14:8A:BC:DE
3. FF:FF:FF:FF:FF:FF
4. 00:50:56:C0:00:08

Step-by-Step Solution: Inspect the first byte (two hexadecimal characters) of each address:

1. E2: In binary, $\text{E} = 1110_2$, $2 = 0010_2 \implies \text{E2} = 1110001\mathbf{0}_2$.
   • LSB is 0 $\implies$ Unicast Address.
2. 4B: In binary, $4 = 0100_2$, $\text{B} = 1011_2 \implies \text{4B} = 0100101\mathbf{1}_2$.
   • LSB is 1 $\implies$ Multicast Address.
3. FF:FF:FF:FF:FF:FF: All 48 bits are 1s $\implies$ Broadcast Address.
4. 00: In binary, $00 = 0000000\mathbf{0}_2$.
   • LSB is 0 $\implies$ Unicast Address.

---

🎯 Exam & Interview Pitfall Check

Core Conceptual Questions

• Are the Preamble and SFD included in the 64-byte minimum frame size? No. The standard 64-to-1518 byte frame boundary begins at the first byte of the Destination MAC address and ends at the last byte of the CRC-32 checksum. The 7-byte Preamble and 1-byte SFD are physical-layer framing constructs stripped by the PHY transceiver before passing the frame to the MAC layer.
• Why can a Source MAC address never be a Multicast address? Multicast represents a group identifier for packet reception. A physical transmission on a link always originates from a single, physical hardware transmitter. Therefore, the Source MAC address must always identify that specific individual hardware interface (Unicast).

Common Interview Traps

• Trap: Confusing Baud Rate with Bit Rate in Manchester Encoding. In Manchester encoding, every bit requires two signal transitions. Thus, a $10\text{ Mbps}$ Ethernet link operates at $20\text{ Megabaud}$. The baud rate is twice the bit rate, not half.
• Trap: Forgetting the difference between MTU and Maximum Frame Size. MTU (Maximum Transmission Unit) refers strictly to the payload size ($1500\text{ bytes}$). The Maximum Frame Size includes the 18 bytes of header and FCS ($1518\text{ bytes}$, or $1522\text{ bytes}$ with VLAN tag).

---

🗺️View Complete Computer Networks Roadmap & Syllabus

← Previous Topic3.2 CSMA, CSMA/CD & Exponential Backoff AlgorithmNext Topic →3.4 Controlled Access & Channelization: Token Ring, FDMA, TDMA, CDMA