Ethernet (IEEE 802.3) Header Formats and the Need for Padding

Ethernet over IEEE 802.3 uses a MAC (Media Access Control) frame that starts with an 802.3 header (destination/source addresses plus a field that is interpreted as either a length or a type, depending on the framing variant) and ends with a Frame Check Sequence (FCS) for error detection. keywordkeywordkeywordkeyword The exact interpretation of the “length/type” field distinguishes Ethernet II from IEEE 802.3 + LLC/SNAP framing.
A key practical requirement is that Ethernet must meet the minimum frame size on the wire; if the payload (and headers) are too small, transmitters add padding so the receiver can reliably detect and process frames. 2

Footnotes

1. IEEE 802.3 / Ethernet framing overview (Length/Type interpretation) - https://en.wikipedia.org/wiki/Ethernet_frame#Ethernet_II_frame ↩

2. Minimum frame size and padding on Ethernet - https://en.wikipedia.org/wiki/Ethernet_frame#Frame_format ↩

1) 802.3 header formats: “length/type” and the two common interpretations

At a high level, the early part of an Ethernet/802.3 frame is standardized: Destination Address (DA), Source Address (SA), and a 2-byte field that is interpreted differently:

• Ethernet II framing: the 2-byte field is an EtherType (identifies the upper-layer protocol, e.g., IPv4).
• IEEE 802.3 framing with LLC (and sometimes SNAP): the 2-byte field is a length (number of bytes in the LLC payload), and the receiver expects LLC headers to identify the network protocol. 2

2) Where LLC/SNAP fits

When the 2-byte field is used as a length, the immediate bytes after it are typically an LLC header (and optionally SNAP) so that multiple higher-layer protocols can be multiplexed over the same link layer format. This is why 802.3 “header formats” are often taught as “MAC header + length/type interpretation + (LLC/SNAP vs not).” 2

Footnotes

1. IEEE 802.3 / Ethernet framing overview (Length/Type interpretation) - https://en.wikipedia.org/wiki/Ethernet_frame#Ethernet_II_frame ↩

2. IEEE 802.3 frame structure and LLC usage (length vs EtherType concepts) - https://en.wikipedia.org/wiki/IEEE_802.3#Frame_format ↩ ↩²

3. LLC and SNAP overview in IEEE 802 networks - https://en.wikipedia.org/wiki/Logical_Link_Control ↩

3) Field-by-field view of the on-the-wire MAC frame

The following conceptual breakdown focuses on the MAC-layer portion (and indicates where LLC/SNAP may appear). Preamble/SFD are part of physical-layer framing and are often excluded from “Ethernet frame size” calculations, but they are present on the wire. 2

Footnotes

1. IEEE 802.3 / Ethernet framing overview (Length/Type interpretation) - https://en.wikipedia.org/wiki/Ethernet_frame#Ethernet_II_frame ↩

2. Minimum frame size and padding on Ethernet - https://en.wikipedia.org/wiki/Ethernet_frame#Frame_format ↩

Why padding is required (minimum frame size)

Ethernet requires a minimum frame size so that the physical-layer signaling and link-layer operation behave correctly for the worst-case conditions on the medium (e.g., collision detection timing on shared media / and consistent minimum transmission time). If the MAC header plus payload (and LLC/SNAP when present) do not reach that minimum, the transmitter adds padding bytes between the payload and FCS. 2

Key intuition

• The MAC must transmit frames long enough to satisfy the Ethernet timing/format constraints.
• If the payload is too short, the frame would otherwise be “too small,” so the MAC pads it with zeros (or equivalent fill) up to the minimum size.
• Receivers use the frame’s length/type and protocol expectations to know how much of the payload is meaningful; padding is ignored at higher layers.

Footnotes

1. IEEE 802.3 / Ethernet framing overview (Length/Type interpretation) - https://en.wikipedia.org/wiki/Ethernet_frame#Ethernet_II_frame ↩

2. Minimum frame size and padding on Ethernet - https://en.wikipedia.org/wiki/Ethernet_frame#Frame_format ↩ ↩²